The United Nations observes the International Day of Women and Girls in Science on February 11 every year to highlight the importance of equal access to and participation in Science, Technology, Engineering, and Mathematics (STEM). The day serves as a powerful reminder that gender equality in STEM is not just a goal but a critical factor in shaping a more inclusive and innovative world.

While India boasts the highest percentage of female STEM graduates in the world – nearly 40% in 2021-22 – only 14% of STEM jobs in the country are held by women. This highlights a critical issue in the tech industry – the ‘leaky pipeline’. Women often exit the workforce at key life stages, including marriage or childbirth, and, sometimes, when family responsibilities intensify. This has led to the underrepresentation of women in leadership roles, innovation, and research.

A study by BiasWatchIndia highlighted the gender disparities within STEM fields. Engineering, for instance, has the widest gender gap, with women representing only 9.2% of faculty members. Fields like biology see slightly higher female representation, at 25.5%. The overall gender gap, however, remains a significant challenge. 

This trend underscores the systemic barriers women face, especially in traditionally male-dominated disciplines such as engineering, physics, and computer science. One question remains: how can women in tech achieve a sustainable work-life balance while overcoming these challenges?

What Women Leaders Think

Women leaders in tech are at the forefront of driving change, pushing for inclusive cultures, and redefining work-life balance. Here are a few stories that provide insights into how women can thrive professionally while maintaining personal well-being.

In a conversation with AIM, Suprabha Mysore, vice president of software development at TransUnion Global Capability Centre (GCC) in India, emphasised the importance of balance for true success. “We can never be great at our work if work is the only thing that we do! True success lies in finding a balance between all the things that matter to us,” she said. 

Suprabha’s perspective highlights that success is not just about career achievements but also about nurturing other aspects of life, such as relationships, health, and personal happiness.

Birlasoft’s chief operating officer, Selvakumaran Mannappan, shares a similar view. He said that work-life balance remains a critical challenge for women in tech, often impeding career growth and retention. Mannappan advocates for workplaces that prioritise flexibility, mentorship, and career development. 

“True inclusion means women in tech can lead, innovate, and grow without sacrificing work-life balance, aspirations, or opportunities,” he stated.

Birlasoft’s BEmpowered Women Leadership program is one such initiative. It equips women with the skills they need to progress into senior roles and fosters a culture where women can balance their personal and professional aspirations.

Kavitha Krishnan, vice president of AI innovations in procurement at SAP Labs India, also stressed the importance of intentional focus and balance. For her, mornings are sacred, a time for deep focus and high-impact tasks. 

“Work-life balance is about making intentional choices,” she explained. For Kavitha, maintaining a focus on health through exercise, mindful eating, and proper rest forms the foundation of her well-being, allowing her to tackle both work and personal life with energy and creativity.

Medhavi Singh, country head at Criteo India, told AIM that the conversation about work-life balance goes beyond flexible hours or remote work arrangements. “True leadership in tech is about prioritising the well-being of those who make business success possible.”

At Criteo, Medhavi emphasised that a supportive culture allows women to excel in their careers while maintaining personal well-being. Whether it’s through adjusting work schedules or providing opportunities for personal growth, creating a flexible work environment is key to empowering women.

Meanwhile, Rekha Sahay Ghosh, head of marketing and communications at Sasken Technologies, said, “It’s about prioritising tasks, being flexible, and setting clear boundaries.”

Efforts Towards Gender Diversity

Efforts to close the gender gap in engineering have been propelled by women like Sudha Murthy, who became the first woman in India to study mechanical engineering, paving the way for future generations.

The Infosys Foundation has taken steps toward increasing gender diversity by partnering with the ICT Academy of Tamil Nadu. The initiative aims to establish Centres of Excellence for Women and Youth Empowerment across 450 colleges in India, offering skills training and job placement assistance to rural learners. The program also includes soft skills development and coding practice sessions to ensure participants are prepared for real-world tech industry challenges.

Similarly, TalentSprint’s Women Engineers (WE) program, supported by Google, is creating an inclusive ecosystem for women in tech. The program has attracted over 1.3 lakh applicants from diverse backgrounds, offering scholarships, mentorship, and career opportunities. With 39% of participants from low-income families and 25% from rural areas, WE aims to empower aspiring women engineers and support their growth in high-growth tech careers across India.

The post How Women in Tech Navigate Work-Life Balance & Career Growth appeared first on Analytics India Magazine.

Content management has come a long way since the early days of the internet. In the early 2000s, websites were the sole digital channel, and managing content was a straightforward but rigid process. Then came the social media revolution, followed by the android boom between 2008 and 2009. However, the real game-changer was the rise of cloud software, which completely transformed how content is created, delivered and managed. 

Traditionally, launching a website meant installing a content management system (CMS), where content creation and presentation were tightly linked. Eventually, as digital landscapes evolved, businesses needed more flexibility. This is when headless CMS, a modern solution that separates content creation from its presentation, was introduced.

Unlike traditional CMS platforms, headless CMS allows developers to use APIs to distribute content seamlessly across websites, mobile apps, and other digital platforms. This separation of content and design not only empowers teams to work independently but also speeds up content updates, enhances user experience, and ensures consistency across multiple channels. 

Role of AI in Headless CMS

Artificial intelligence (AI) plays a significant role in content management. Nishant Patel, founder and CTO at Contentstack, said, “AI today is largely centred around generative AI, which excels at content generation. Many marketing websites and mobile apps leverage AI tools built into our software to create dynamic content effortlessly.”

One of Contentstack’s AI-driven innovations is Brand Kit, a tool designed to capture a brand’s unique voice and tone. When companies use generative AI within their CMS, they ensure that the content aligns with their identity, which makes messaging more authentic and impactful.

Another breakthrough product, Automate, is changing the game in content integration. Patel shared a compelling use case from Golfbreaks, a travel company specialising in golf vacation packages. 

Previously, compiling these packages was a manual, time-consuming task that took days. By integrating AI-powered LLMs like Llama, Gemini, and OpenAI into Automate, Golfbreaks reduced this process to mere hours. The system curates deals, structures them in a suitable format, uploads them to its website, and prepares them for approval – all in record time.

Competition in the Field 

When it comes to the global adoption of headless CMS, North America is leading the charge. The US and Canada have been at the forefront because of their robust digital infrastructure, a thriving tech ecosystem, and a high concentration of businesses seeking agile content management solutions.

Sectors like e-commerce, media, and technology have embraced headless CMS to deliver seamless content experiences across multiple platforms. This early adoption has spurred continuous innovation, with both startups and established players driving the evolution of headless CMS solutions.

Speaking about competition, Patel said, “We focus on mid-sized companies and large enterprises like Fortune 2000 companies and major brands around the world that are using Contentstack to deliver their content. This includes websites, mobile apps, gaming software, and more.”

He mentioned that in the traditional content management space, there are incumbents like Adobe, Sitecore, and Optimizely, and they have recently started making investments in headless CMS. 

What’s Next?

The global headless CMS software market is on the rise. Valued at $0.71 billion in 2023, it’s set to skyrocket to $3.81 billion by 2032, growing at a CAGR of 20.5%, according to Business Research Insights.

The surge is driven by businesses prioritising SEO, performance optimisation, and omnichannel content distribution. Headless CMS is becoming the backbone of modern digital experiences. The integration of AI, automation, and cloud-based collaboration is reshaping content management, making it more dynamic and future-proof.

For developers, marketers, and business leaders, adopting a headless CMS is more than just an upgrade; it could be a transformative shift. 

The post The Future of Content is ‘Headless’ appeared first on Analytics India Magazine.

Would you rather work from home or go to the office every day? Just a few years ago, employees had more control over this decision. But today, the choice is increasingly out of their hands as companies rethink work structures, balancing between hybrid and fully on-site models. 

The debate over remote versus in-office work continues to intensify. While some companies are leaning into flexible models, others are calling their workforce back. 

JPMorgan, for instance, recently announced that employees must return to the office five days a week starting in March 2025, ending its pandemic-era hybrid policy. A memo from the leadership acknowledged employee preferences but stated firmly, “We think it is the best way to run the company.”

This shift isn’t limited to Wall Street, India’s tech and IT companies are also refining their work models. Accenture, for example, has been allowing employees in Bengaluru to visit the office once every three months. However, starting February 2024, they will now be required to come in once a week. 

Infosys has also moved towards a more structured hybrid policy. By late 2023, the company mandated that employees must work in the office at least 10 days per month, making full-time remote work an increasingly rare option in 2025. 

Some leaders strongly advocate for in-person collaboration.

Rizwan Khan, VP of people and culture at Porter, told AIM, “At Porter, we believe direct, in-person interactions foster innovation, mentorship, and strong team bonds. On-site exchanges at the workplace drive employee motivation, enable mentorship, and promote knowledge sharing, all of which contribute to individual and collective growth.” 

He added that while the pandemic highlighted the challenges of fully replicating these benefits in a remote work environment, they also recognise the importance of flexibility for their employees. 

To support the team, the company has implemented a set number of work-from-home days annually and remains open to accommodating additional flexibility based on genuine needs.

Balancing Flexibility and Productivity 

The State of Recruitment Trends (SORT) survey provides insights into how hybrid work impacts businesses. Recruiters rate hybrid productivity at 6 out of 10, suggesting that while effective, it may not be optimal for all companies. 

At the same time, 73% of employees cite work-life balance as the most valued aspect of their work environment, reinforcing the idea that flexibility, when managed well, can support both well-being and efficiency. 

In an exclusive interaction with AIM, Venkataramani Suresh, CEO, foundit, mentioned, “Employee retention and satisfaction are paramount in today’s competitive talent market. Companies are under pressure to balance flexibility with productivity, creating work environments that meet the diverse needs of their workforce. 

“As businesses explore hybrid structures, the ability to manage distributed teams effectively while ensuring alignment with organisational goals remains a key challenge.” 

He further explained that the financial considerations of office space utilisation cannot be overlooked. Hybrid models offer potential cost savings, particularly as companies optimise their real estate footprints. However, recent trends suggest a subtle shift in the market. 

Many companies are optimising their office space costs, but a subtle market shift is underway. Hybrid roles have declined by 9% in the past six months, while in-office positions have risen by 7%.

The conversation about work models is complex, and there is no one-size-fits-all answer. Arvind Krishnan, president and CEO (wellness) at Happiest Health, believes that work preferences are subjective. While some find hybrid arrangements perfect, others are more productive working entirely from home. 

“Ultimately, someone has to pay for the work I do, so the nature of the job dictates what’s feasible. For example, if I were a surgeon, I wouldn’t say that I’d like to work from home; it just wouldn’t work! Some jobs simply require being on-site,” he said. 

What’s Next?

The COVID-19 pandemic forced companies to reevaluate office leases as remote work led to increased office vacancies. By 2022, India’s office vacancy rate hit 16.4%, with major hubs like Bengaluru and Mumbai witnessing sharp declines in occupancy. Now, businesses are recalibrating their strategies. 

The 2024 SORT survey revealed that 33% of organisations have fully adopted a hybrid model, with another 9% in transition. Co-working spaces have also surged in popularity, with a 42% year-on-year increase, reflecting both cost-optimisation efforts and evolving employee preferences.

As companies like Infosys, Wipro, and Google India maintain hybrid models, others like TCS and Amazon are steering back to full-time office setups. foundit’s survey highlighted that 47% of companies continue to operate with a traditional five-day office schedule, while 33% have embraced hybrid models, 9% are transitioning, and 10% have gone fully remote. 

While the debate continues, one thing is clear, the future of work isn’t about choosing between home and office, but about striking the right balance between flexibility, productivity, and business needs.

The post What’s Pushing Companies to Choose On-Site Work Over Hybrid? appeared first on Analytics India Magazine.

Tools like Cursor, Windsurf, and even Claude have taken over the coding world since last year, offering auto-completion capabilities and generating code for most use cases. While Indian companies were initially hesitant to use them, their use is now becoming more accepted, and there is a shift towards in-house development instead of solely relying on existing market solutions.

In a recent Reddit discussion, a full-stack developer working in a product-based firm highlighted how the teams have been using AI tools extensively and have also built an in-house extension of VS Code to integrate into ChatGPT, Gemini, and Claude. The tool also has access to the company’s entire codebase. 

“To use it, you simply enter the ticket number you’re working on in the extension’s chat box. The AI retrieves the relevant Jira ticket, analyses its details, and generates the required code,” the developer explained, adding that the tool can do close to 60% of the work. Later, the developers can review and edit the rest of the code. 

The team is also actively updating the extension to increase the productivity of its employees with the aim of completing at least 80% of the tasks using AI. “It’s not replacing me; it’s enabling me to concentrate on the most important pieces while taking care of the repetitive tasks.”

In the discussion, developers expressed optimism about the tool while also expressing fear and criticism about how such tools could threaten developers’ jobs. Some users said if AI is able to write the code for the software that the firm is building, the company must be building low-level stuff and might be at threat of closing down soon. 

Startups Leading the Race

According to a recent survey by GitHub, 56% of Indian developers said that using AI tools helps them boost their chances for employment owing to the skills they develop. Moreover, around 80% of them believe AI tools have improved code quality.

While startups are eagerly embracing these tools to accelerate development, legacy enterprises remain hesitant due to cost and implementation concerns. Abhishek Upperwal, CEO of Soket AI Labs, told AIM that these tools have become indispensable for him. 

“We’ve been using AI coding tools at Soket AI Labs for a while now – both open and closed source. Honestly, the whole experience depends on the backend model. Tools like Claude 3.5 Sonnet and GPT-4o are great for easy stuff like refactoring, but if you’re looking at complex code generation, O1 is the real deal,” Upperwal said.

However, AI tools still have limitations. “None of these tools are useful when it comes to things like CUDA kernel optimisations – we’ve been working with Triton [and are] still waiting for AI to start building AI,” he added.

While interacting with AIM, Yulu CTO Naveen Dachuri revealed that his team has been experimenting with Claude and Windsurf for many coding tasks and has witnessed good results when it comes to shipping faster.

While startups are leveraging AI coding tools for rapid prototyping and development, large enterprises remain cautious. The primary roadblock? Cost.

According to Upperwal, even if productivity shoots up tenfold, a $20 license for 5 lakh employees is not feasible even for enterprises.

Adarsh Shirawalmath from Tensoic and Adithya S Kolavi from CognitiveLab echoed similar thoughts about the cost of such models. 

What are Some Favourite Tools?

Dharmesh Shah, CTO at HubSpot, recently shared his experience on LinkedIn while helping his 14-year-old son choose a coding tool for an AI-based web application.

“I personally use Cursor for my development, but [I’m] not sure if it’s the best choice for a ‘generate an entire app’ use case. I’ve heard good things about Bolt and Windsurf,” he said, adding that he is also a fan of Replit.

Responding to Shah, Jonathan Tushman, CPO and CTO at Hi Marley, said, “Right now, it’s a toss-up between Cursor and Windsurf. But for beginners, Windsurf has the advantage. They’ve just released a major upgrade that includes web search, which is a big deal.”

“The biggest thing for your 14-year-old to learn is Git and GitHub. Keeping a clean branch and working with PRs (pull requests) is key. Without that, it can be a frustrating experience,” Tushman added.

A structured approach is crucial for developers working with AI-generated code. Mike Beech, an early-stage startup tech expert, revealed a technique. “First, enter your requirements and ask for a more detailed specification. Then, get a step-by-step breakdown as a markdown to-do list. This ensures the AI stays on track and avoids unnecessary complexity.”

While AI coding tools are undoubtedly accelerating software development, they are not a substitute for fundamental programming skills. 

“These tools are only as good as the person using them. If you know your stuff, they’ll blow your mind with the quality. But hand them to someone without decent programming skills, and it’s chaos. They’re powerful, but they’re not magic. Yet,” Upperwal pointed out.

The post AI is Taking Over Coding at Indian Companies appeared first on Analytics India Magazine.

While many believe India has missed the GenAI bus, jumping on Large Concept Models (LCMs), the next emerging technology, could present a unique opportunity for the country to explore.

LCM technology, still at its nascent stage, exhibits characteristics that could help solve complex problems and be a game-changer for India to lead the AI race.

But is this the right move? Should India invest in LCMs? Is it a bad idea? Do experts believe there is a better option?

The LLM Race is Tough Enough

Even though India is leading the charge in AI agents, it is incredibly tough to compete for a leading spot in terms of a foundational model as we know it.

Solutions like AI4Bharat and Ola’s Krutrim, among others, have made headlines. Several startups in India are working on AI-powered solutions to automate mundane tasks and have been doing exceptionally well.

Despite all this, India is not leading the race, and the reason is lack of capital. Mohandas Pai, head of Aarin Capital and former CFO of Infosys, spoke to AIM last year about this when asked about the lack of innovation from Indian IT. “Who will give $200 million to a startup in India to build an LLM?” he wondered.

“Why is nothing like Mistral coming from India? There is nobody…Creating an LLM or a big AI model requires large capital, time, a huge computing facility, and a market. All of which India does not have,” he further said.

At a quick glance, many leading LLMs stand out, with countless others joining the list every other day. With India entering the AI race, it may not be an outright “aha moment.”

Language Concept Models Are Unexplored, But Is It Useful?

Large Concept Models (LCMs) are in the nascent stages of development. However, as the research paper by Meta explains, they could have immense benefits for India.

LCMs are language and modality agnostic, meaning they could work well for all diverse language and cultural complexes in India without special efforts. It uses a sentence embedding space, SONAR (Sentence-Level Multimodal and Language-Agnostic Representations), which supports 200 languages, including Hindi, Kannada, and more.

The architecture behind an LCM allows it to tackle complex problems (advanced reasoning) better than LLMs, as it emphasises semantic understanding, hierarchical processing, reasoning, and planning.

It also supports a larger context length by default because the smallest unit here is a “sentence” or a concept.

They are also excellent at document summarisation and interactive content creation, as explained in ADaSci’s blog titled ‘A Deep Dive into Large Concept Models (LCMs)’.

So, by shifting their focus from LLMs to LCMs, Indian companies could choose to eliminate putting efforts into Indic language models, discover a new way to reduce computational costs, and tackle complex problems in the country.

Will This Help India in the AI Race?

While the tech speaks for itself, India recently received a significant boost with a ₹2,000 crore investment in Budget 2025 for the IndiaAI mission. 

With this in mind, AIM spoke to Amlan Panigrahi, a GenAI engineer at Deloitte, about this idea. “I feel rather than investing in LCMs, which is an initiative of Meta, India should focus on investing in existing Sovereign AI initiatives to solve the nation’s problems first. Entities like Sarvam.AI, OdiaLlama and Indic language models should be targeted more to strengthen the foundation of the country in AI,” he said.

He also mentioned that exploring Graph Transformers could be a novel and impactful research direction, as it has the potential to solve contemporary industry problems.

Bilal Yoosuf, a senior consultant in data science and engineering at TNP India, spoke to AIM and shared his thoughts, which resonated with a similar sentiment, “India still has a huge opportunity to explore LLMs, especially in the social sector, before shifting focus to LCMs.”

“A major part of the population hasn’t even accessed GenAI yet. With the government’s new AI investments, prioritising LLMs for regional languages, education, healthcare, and governance could create a real impact. Once we build strong AI solutions for our own people, we can then look at leading the next wave of AI innovation globally,” Yoosuf added.

On the other side, experts like Utsav Khandelwal from Maxim AI believe that by investing early in LCM research, India could position itself as a leader in next-gen AI and mirror its success with cost-effective space missions like ISRO’s.

LCM is an uncharted territory for most of the world. It remains to be seen when an AI researcher from India picks it up and explores what it can do for India.

The post Should India Shift its Focus from LLMs to Large Concept Models? appeared first on Analytics India Magazine.

French artificial intelligence startup Mistral has announced plans to invest “several billion euros” in building its first data centre in France. The company aims to gain full control over data storage and processing power.

Mistral co-founder and CEO Arthur Mensch made the announcement just before the AI summit in Paris, scheduled for February 10 and 11, 2025. At the summit, world leaders and tech industry executives will discuss the future of artificial intelligence.

According to Mensch, Mistral’s goal is to manage the entire AI value chain, from hardware infrastructure to software. 

The move aligns with France’s push to position itself as a prime data centre hub, employing its low-carbon nuclear energy and readily available development sites to attract major investors.

The French AI lab introduced its AI assistant, Le Chat, last week. 

President Emmanuel Macron expressed his support on X, exclaiming, “Vive Le Chat!” The government has backed the startup with fresh contracts, striking deals with the Ministry of the Armed Forces and the public employment agency, France Travail.

Silicon Valley-based Cerebras provides computing power. Cerebras, a challenger to Nvidia in AI training, is focused on inference, delivering AI responses efficiently. 

Mistral is also preparing for an initial public offering (IPO), Mensch announced earlier during an interview with Bloomberg at the World Economic Forum (WEF). 

“We’re not for sale,” Mensch told Bloomberg, adding that the company plans to open a new office in Singapore to strengthen its presence in the booming Asia-Pacific market. It also expands operations across Europe and the United States, building on its mission to compete with industry leaders like OpenAI.

The post Mistral AI Lab to Build First Data Centre in France  appeared first on Analytics India Magazine.

The last day of MLDS 2025, India’s largest developer conference, kicked off with Paras Chopra, founder of Turing’s Dream (now LossFunk), discussing how India can build advanced AI models for its own needs. 

Chopra stressed the urgency of AI investment, with the U.S. pouring $500 billion and China $150 billion into the field.“Does it make sense for the world’s most populous country to be left behind permanently?” he asked. 

To stay competitive, India must develop foundation models, drawing on DeepMind’s breakthroughs and the country’s strong AI talent. He stressed the need for India to have home-grown superintelligent AI models.

His vision for India competing in the AI SOTA landscape was clear—an “ISRO of AI” by 2025, making India a global AI powerhouse.

Building on the theme of AI’s impact on the developer ecosystem, Invact Metaversity CEO Tanay Pratap explored AI’s impact on programming, presenting both challenges and opportunities for developers. 

Drawing on his experience at Microsoft and Cisco, where he trained over 1,000 students in coding, he noted that while AI can predict tokens, it lacks a true understanding of code functionality, highlighting the need for human oversight and improved training models.

Taking this discussion further into real-world applications, Abhishek A, the chief technology officer at SmartQ, and Keshav Meda, co-founder and chief growth officer, presented their insights into how SmartQ is acting as a bridge between humans and machines, enabling structured communication. 

They highlighted the importance of purpose-built hardware, software optimisation, and high-quality data in accelerating this shift.

During the talk, Abhishek showcased Leva, Smart Cube’s in-house AI tool that automates software development, improving overall productivity. Looking ahead, he predicts a future where AI writes code at lightning speed, customises user interfaces, and manages real-time updates, fundamentally reshaping the way software is built and maintained.

Shifting the focus to AI’s role in financial services, the event also saw Mathangi Sri Ramachandran, chief data officer at Yubi, lead a workshop on how Large Language Models (LLMs) are transforming finance. She spoke about their role in risk management, fraud detection, automated reporting, and personalised customer experiences, providing real-world applications and fine-tuning strategies.

As AI continues to disrupt industries, its convergence with Web3 is unlocking new possibilities. Vivek Gupta, CTO at CoinDCX, shared insights on how Web3 and AI are shaping the next wave of innovation in his session on building crypto in an AI world. 

With AI playing a more central role in decision-making, ensuring strong governance is crucial. Mantra Labs CTO Kumar Sambhav Singh talked about the importance of strong data and AI governance, highlighting the need for accuracy, reliability, accessibility, and compliance with industry standards. 

In addition to this, Saurabh Pramanick, data governance officer at Bank Muscat, highlighted executives driving AI governance in organisations with larger budgets and the urgency of their investments. The session covered the forecast for new AI models, the tools and frameworks in use, and confidence in AI capabilities. It also examined key AI risks, regulations, and governance ownership.

AI is also transforming the education industry. Navin Reddy, founder of Telusko Edutech, talked about how AI is changing the way students learn and solve problems. While they once turned to books, peers, and online trainers, many now rely on AI for debugging and troubleshooting. He stressed that educators now have an even bigger role in helping students use AI effectively while building a strong foundation in core concepts.

As MLDS 2025 wraps up, one thing is clear, AI isn’t just something we use; it’s shaping the way we work, create, and solve problems. The recent disruption by DeepSeek has helped shift perspectives, turning curiosity into confidence and scepticism into acceptance. 

Register Now

Date: Thursday, February 13, 2025

Time: 4.00 PM

Location: SmartQ Bengaluru Office

Save the date – Don’t miss out! Seize this opportunity to become a 10X smarter engineer.

The post MLDS 2025: Key Highlights from  Day 3 appeared first on Analytics India Magazine.

The AI-driven solutions developed for enterprise-scale use are enhancing customer experiences, reducing call handling times, and automating millions of daily interactions across industries such as banking, insurance, and retail.

In a presentation at MLDS 2025, AI innovators at Gnani ai showcased how AI-powered voice agents they developed are transforming industries.

Ganesh Gopalan, CEO and co-founder of Gnani.ai, shared an example where an AI voice agent helped a user who had forgotten about their PAN card application within just 24 hours. 

Voice Agents That Talk, Think, and Assist in Real Time

AI voice agents are more than just chatbots. “They are the next level of automation—handling real-time conversations at scale,” Gopalan explained.

Their AI-powered voice customer service agents have already processed over 30,000 concurrent calls and millions of daily interactions.

The technology works across platforms, including telephone, WhatsApp, and iMessage, automating customer service and reducing operational costs.

“Our AI can lower call handling time by 15% and improve outcomes for customer support agents by 40%,” Gopalan shared.

The Challenges of Building AI Voice Agents

Creating an AI voice assistant is not as simple as using a chatbot. Bharat Shankar, co-founder and chief product officer at Gnani.ai, highlighted the need for high-quality speech recognition that works across various accents and languages. 

Another major challenge is handling interruptions naturally, as customers frequently cut in during conversations. “Customers interrupt a lot. AI must process interruptions smoothly,” he said. 

Also, reducing response time is also critical since long pauses negatively impact user experience. “Long pauses kill the experience. Our AI responds in under two seconds,” Shankar noted. 

While data security remains a top priority, especially for enterprises that require strict compliance, scaling AI-powered conversations requires advanced cloud infrastructure to handle millions of conversations daily. 

AI That Thinks and Speaks Instantly

To solve these challenges, Gnani.ai has developed a “speech-to-speech LLM” (Large Language Model) that processes both audio and text simultaneously for faster, more efficient interactions.

 “This is the future—AI that thinks and speaks instantly,” Shankar said, revealing that the technology will be showcased at NVIDIA’s GTC conference in California in March.

In the insurance sector, a live demonstration was showcased where an AI-powered assistant seamlessly handled customer queries by greeting callers politely, switching languages from English to Hindi mid-conversation, recognising customer preferences, and offering customised insurance plans. 

The entire interaction was completed in under two minutes, including a WhatsApp link for document submission. In banking and government services, AI voice agents are helping streamline processes, such as retrieving PAN card details in record time, demonstrating how AI can improve government services. “Customers no longer need to wait days for a simple query—AI speeds up the process,” said Gopalan.

Agent Builder Platform

To make AI adoption easier, Gnani.ai launched an “Agent Builder” platform, an end-to-end tool for businesses to create, test, and deploy AI-powered voice agents. The platform offers support for 40+ languages, 400+ pre-built AI templates, real-time voice orchestration, and seamless CRM and enterprise integrations. 

“With one click, companies can deploy AI agents and automate conversations instantly,” said Shankar. 

Looking ahead, AI voice agents will understand emotions and adjust their tone, seamlessly switch between multiple languages, and handle complex multi-step tasks automatically. 

As businesses increasingly rely on AI to reduce wait times, improve accuracy, and enhance human-like interactions, early adopters will gain a significant competitive edge. “The future is here. AI voice agents are not just changing businesses—they are transforming customer experiences forever,” Gopalan concluded.

The post Gnani.ai Showcases Real-Time AI Voice Automation at MLDS 2025 appeared first on Analytics India Magazine.

Hallucinations continue to be one of the most critical challenges with AI. Although databases and a few other methods can mitigate this, they’re not the only solution. Some AI frameworks, like CTGT, have aimed to nearly eradicate hallucinations, but their effectiveness has yet to be universally felt. Notably, this approach involves another form of AI.

At MLDS 2025, India’s largest GenAI summit for developers organised by AIM, Ratnesh Singh Parihar, principal architect at Talentica Software, said his team uses AI to fix AI hallucinations. 

Parihar discussed the difficulty of handling over 10 million stock-keeping units (SKUs) in e-commerce search with AI and how AI is used to fix AI-related problems.

AI Fixing AI Hallucinations

Ritesh Agarwal, solution architect at Talentica Software, explained that when users search for items, such as ‘pink t-shirt for toddlers’, traditional AI methods convert queries into embeddings and use cosine similarity to find relevant products.

However, hallucinations occur when irrelevant items like jeans or wristbands appear in search results. To combat this, his team integrated AI-powered validation checks using OpenAI to flag inaccurate results that stemmed from a hallucination with inconsistencies generated by AI tools.

According to Agarwal, Talentica Software ran test queries to retrieve results from its comparison models, whether based on semantics or cosine similarity. The system provided a simple true or false flag, which they stored in our database.

Parihar further expanded on this idea, “Let’s say you want to generate a research paper using ChatGPT. It will generate good content, but the conferences will reject it because they can figure out what you say is machine-generated. No? But, some other AI tools can take that content and humanise it. So, AI has generated one problem, but you can use another AI tool to solve it. That’s how you can go about it.”

Minimising Humans in Building AI

While he mentioned that it is better to build things manually unless you absolutely need to, Parihar also stressed the need to minimise humans when building AI. “You cannot say…’I will create XYZ’, and then some people will come and verify the XYZ. You need to build the bots.”

“You might need a person who knows a lot of AI tools. So they can use the tools. And, most importantly, you require someone who can convert those tools into bots, human-like,” he added.

Combining AI Tools Not Just to Fix Problems But To Save Cost

The team at Talentica Software mentioned specifically that they used both Llama3 and OpenAI to combine with one another for better results, maintain efficiency and reduce the percentage of errors.

In particular, Llama 3 was used for large-scale product categorisation and tagging, significantly reducing costs compared to OpenAI, according to Parihar. 

In combination, OpenAI was used as a validation mechanism to identify errors and hallucinations in search results for the e-commerce system. The AI system compared search results against image descriptions and product details, flagging inaccuracies through a true/false validation system. 

Parihar also provided insights into why they switched to Llama 3 for tag or category generation and how they saved 97% in the cost required when dealing with SQL queries. With OpenAI, it cost them around $500 for 1 million SQLs and, in comparison, the Llama 3 model cost them just $15 for the same.

It is intriguing to see that the AI we built comes in handy in solving errors generated by the same family and reducing costs. We have come a long way in a short span of time, from relying on HTML to reduce hallucination overdose to trusting another AI. Fortunately, we do not need extra human efforts to find solutions to problems, ultimately making good use of AI.

The post Using AI to Fix Problems It Creates appeared first on Analytics India Magazine.

Day two of MLDS 2025, India’s biggest GenAI summit for developers, hosted by AIM Media House, continued with just as much energy, excitement, and insightful discussions as the first day. Multiple tech enthusiasts attended the event.

Today’s captivating talk featured AI innovators Ganesh Gopalan, CEO and co-founder of Gnani AI, and Bharat Shankar, co-founder and chief product and engineering officer at Gnani AI, highlighting the transformative impact of AI-powered voice agents across various industries.

A standout example mentioned was a user who forgot about his PAN card application and was assisted by an AI voice agent in just 24 hours. 

These AI-driven solutions, built for enterprise-scale use, are enhancing customer experiences, reducing call handling times, and automating millions of daily interactions across industries like banking, insurance, and retail.

AI voice agents are more than just chatbots. “They are the next level of automation—handling real-time conversations at scale,” Gopalan explained. Their AI-powered voice customer service agents have already processed over 30,000 concurrent calls and millions of daily interactions.

The technology works across platforms, including telephone, WhatsApp, and iMessage, automating customer service and reducing operational costs.

“Our AI can lower call handling time by 15% and improve outcomes for customer support agents by 40%,” Gopalan shared.

In another interesting session, Ritesh Agarwal, solution architect at Talentica Software, discussed hallucinations as a key challenge with AI. 

Although databases and a few other methods can mitigate this, they are not the only solution. 

Agarwal and his team found it difficult to use AI to handle over 10 million stock-keeping units (SKUs) in e-commerce searches, so they used AI to fix AI-related problems.

He further explained that they use test queries to retrieve results based on semantic or cosine similarity. They then send the image and query to OpenAI for validation of stock items, which returns a simple true or false flag indicating accuracy. This flag is stored in the database to manage hallucinations.

Another highlight of the day was an exciting workshop, “Building Scalable Multi-Agent Systems with Gemini: From Scratch to EV Industry Report,” led by Lavi Nigam, a developer relations engineer at Google Cloud.

This hands-on session delved deep into the world of multi-agent systems, showing participants how to build scalable AI applications from the ground up using Google’s Gemini AI models. Attendees explored key design patterns, architecture, and essential tools needed to develop robust, intelligent AI systems.

Finally, a prominent highlight of the day was the “40 Under 40 Data Scientists” awards at MLDS 2025, hosted by AIM Media House. 

This prestigious recognition brought together some of India’s brightest minds in data science, celebrating their innovation, impact, and contributions to the industry.

These young data scientists are driving the future of analytics in India, shaping the landscape with their vision and expertise. The award highlights real innovators and achievers, setting them apart as leaders in the field. AIM’s expert panel of editors and industry veterans carefully reviewed and selected the winners, making this a truly elite recognition in the world of data science.

The post MLDS 2025: Key Highlights from Day 2 appeared first on Analytics India Magazine.

The AI industry is accelerating rapidly, and this is evident in the introduction and application of AI agents. A few years back, AI was just an LLM wrapper, but now, with options like Operator by OpenAI, it is much broader.

At MLDS 2025, India’s largest GenAI summit for developers, organised by AIM, Siddhant Goswami, co-founder of 100xEngineers, an AI lab specialising in Generative AI, explained how AI agents have evolved and changed over the past few years.

Goswami noted that over the past two years, ever since he started developing AI Agents, their definition has changed significantly each year and each quarter. He referenced “God In A Box,” which many had initially perceived as an AI agent, but it merely was an LLM wrapper. He emphasised that perspectives are shifting and that the entire industry is advancing rapidly.

Not just Goswami, but many Indian founders, in general, love AI agents.

The Phases of AI Agent Evolution

The journey of AI agents can be divided into distinct phases, each introducing a new level of sophistication and functionality. Initially, everyone was clueless about why they existed, but they are now building meaningful solutions based on this knowledge.

Goswami outlines five phases of AI development. The first phase began with large language models (LLMs) like ChatGPT in 2022, which could process text inputs but lacked advanced reasoning. The second phase improved context windows, allowing users to input extensive texts, leading to the practical use of corporate databases.

The third phase introduced retrieval-augmented generation, enabling LLMs to access external knowledge. The fourth phase brought multimodal capabilities, allowing LLMs to process images, voice, and video. In the fifth phase, LLMs utilise memory layers, enhancing their ability to remember and adapt responses.

Building an AI Agent as an Artificial Being

Summing up the phases of building an AI agent, Goswami told AIM, “We are kind of recreating a being, an artificial being. We first gave them hands and legs, then we taught them to learn the language, and then we gave them memories so they could remember things. I feel that’s the evolution of AGI itself.”

With this type of advancement in AI agents, he believes we can expect to achieve AGI sooner.

Especially when Indian founders, in general, love AI agents.

AI Agents Need Human Intervention

At MLDS, Goswami mentioned that AI agents can hallucinate and go off the rails. So, a human-in-the-loop model is needed to monitor the operations.

“Even if you build the best technology right now, without having a human in the loop and giving feedback, I don’t think we would be able to achieve the level of intelligence we are expecting from this.”

Despite the exciting progress, Goswami warned that not everything requires an AI agent. Many businesses rush into agent-based automation without understanding the trade-offs.

Most AI problems can be solved with a simple LLM and a RAG setup. Companies implementing AI solutions should prioritise simplicity. He suggested not to overcomplicate things and only introduce agents if necessary. If a task requires multiple steps, reasoning, and external tools, that is when we should consider using an AI agent.

What’s Next for AI Agents?

It looks like AI agents will become increasingly autonomous, with longer memory retention and real-world action-taking abilities.

Research and Market’s report on ‘AI Agents Market Analysis projects the market for AI agents to grow from $5.1 billion in 2024 to $47.1 billion in 2030, with a CAGR of 44.8% during 2024-2030.

For now, with human observation added, the definition of AI agents should continue to evolve. It’s exciting to see what the future holds for them.

The post As AI Agents Evolve, So Does Their Definition appeared first on Analytics India Magazine.

Day one of MLDS 2025, India’s biggest GenAI summit for developers, hosted by AIM Media House, was a day filled with energy, excitement, and forward-thinking discussions. The day witnessed the presence of close to 1000 attendees.

One of the day’s interesting talks came from Rahul Bhattacharya, AI leader at GDS consulting, EY, who delved into the world of the agentic workforce, the next frontier in AI-driven workspaces. Bhattacharya highlighted how assessing the risks of integrating AI agents into the workforce is just as crucial as measuring the benefits.

He broke down what makes a system truly “agentic.” To be considered an agent, Bhattacharya explained, a system must have the ability to interact with its environment, make decisions based on its observations, and learn from its actions to continuously improve. 

He elaborated, “A key ability is making decisions, where the agent chooses the best action based on set rules, goals, or rewards. Over time, it should learn from past experiences and feedback to improve its performance.” This blend of adaptability and decision-making could redefine how we think about work in the future.

In the digital content world, where video has taken center stage, Arvind Sasikumar, co-founder and CTO at Quinn, shared his insights on the critical importance of video compression. 

Sasikumar emphasised that optimising video transcoding isn’t just about reducing file sizes, but it’s about enhancing the user experience by eliminating buffering, which can be a dealbreaker for viewers. He explained how even a 5% reduction in file size can make all the difference for users with lower internet speeds. 

“Consider this: If 100 users have a 1.9 Mbps connection but the video they’re watching requires a 2 Mbps bitrate, every single one of them will face buffering. But by optimising compression, we can ensure that all 100 users enjoy smooth playback,” Sasikumar said. 

As organisations race to integrate AI into their business strategies, there are hurdles to overcome. Chirag Jain, vice president of AI Practice at Genpact, took the stage to shed light on why businesses must first define their long-term goals before jumping into AI implementation. 

He compared it to the vision of a sports team– just as India’s cricket team aspires to be the best in the world, businesses also require a clear, long-term goal to guide their AI strategy. 

Jain underscored that a strategy grounded in core values, ethical considerations, and regulatory compliance is key to ensuring AI’s success within an organisation. “Nobody likes to see a team playing unfairly,” he said, reinforcing the need for transparency and ethical decision-making in AI adoption.

The day also featured Siddhant Goswami, co-founder at 100xEngineers and Tech Influencer, who spoke on the evolving role of AI agents. While AI agents are becoming increasingly autonomous, Goswami reminded the audience that human intervention is still essential. 

“We can’t completely rely autonomously on them. We need human feedback. These agents are self-directed, capable of planning and executing multiple steps based on feedback and updated goals,” Goswami shared. 

The excitement continued with a workshop by Gopala Dhar, an AI engineer at Google Cloud, and Lavi Nigam, a developer relations engineer at Google Cloud, on building real-time applications with the Gemini Multimodal Live API. This hands-on workshop served as a holistic guide to using Gemini’s multimodal capabilities to create sophisticated applications that can see, hear, and interact naturally. 

Also, the day witnessed several paper presentations, including Life Stage Customer Segmentation by Fine-tuning Large Language Models by Nikita Katyal, Head of Analytics and AI at Central Retail Corporation, and Vivek Vishwas Vichare; A Context-Aware Multi-Agentic Multi-Modal LLM Architecture for Digital Marketing by Vivek Vishwas Vichare, head of data sciences and analytics at Pixis, and more. 

These papers showcased the innovative ways AI and generative models are reshaping industries like retail, digital marketing, and healthcare, demonstrating the immense potential of AI to solve complex problems across various domains.

Day one of MLDS 2025 left attendees with plenty to think about. 

The post MLDS 2025: Key Highlights from Day 1 appeared first on Analytics India Magazine.

In the digital content world, video has become the dominant format, and with that, video compression is more critical than ever. Speaking at India’s Biggest GenAI Summit for developers, MLDS 2025, Arvind Sasikumar, co-founder and CTO at Quinn, shared insights on optimising video transcoding to balance efficiency, quality, and playback performance.

The benefits of video compression are clear, smaller file sizes lead to reduced storage costs and lower data transfer expenses. But beyond these obvious advantages, compression directly impacts user experience. One fundamental aspect that often goes unnoticed is that video buffering is a binary state: a video is either playing smoothly or buffering. There’s no middle ground.

Sasikumar explained, “Consider an example. If 100 users each have a 1.9 Mbps internet connection, but the video they are watching has a 2 Mbps bitrate, every user will experience buffering. However, by reducing the file size by just 5%, all 100 users can enjoy seamless playback. This demonstrates why compression is not just about reducing numbers, it’s about eliminating interruptions that degrade the viewing experience.”

There are two primary types of compression: lossy and lossless. Lossy compression reduces file size by removing some data through predictive algorithms. This is the preferred method for video and audio because human perception can compensate for minor quality losses. Lossless compression retains all original data, ensuring no quality loss. This is ideal for text and data files where precision is crucial but is not practical for video due to high storage requirements.

It’s All About Perception

Humans don’t perceive pixels individually; instead, we process visual information contextually. Compression algorithms leverage this by reducing redundant data without noticeable quality loss. 

Sasikumar mentioned that one common technique is chroma subsampling, where brightness (luminance) is prioritised over color detail. Since the human eye is more sensitive to brightness than color, encoding schemes like 4:2:0 cut data usage by nearly half without significant perceptual impact.

Beyond individual frames, modern video encoding techniques exploit similarities between consecutive frames to achieve higher compression rates. Instead of storing each frame as a separate image, encoders analyse differences between frames and store only the changes.

A robust compression algorithm must optimise three key areas: compressing individual frames efficiently, minimising redundant data between frames, and leveraging human perception to maintain quality at lower bitrates. One of the most effective strategies is motion estimation and motion compensation. 

Instead of encoding pixel-by-pixel differences, encoders track objects as they move across frames and store them as motion vectors. This significantly reduces the amount of data required to represent motion.

He explained with an example, if a ball moves across the screen while the background remains static, the algorithm records only the ball’s movement rather than re-encoding the entire frame. This principle underpins most modern video encoding formats. However, the accuracy of motion estimation affects compression efficiency. 

Role of I-Frames in Video Encoding

I-frames (Intra-coded frames) are key reference points in video encoding. Their placement plays a crucial role in ensuring smooth playback and efficient compression. The first frame should be an I-frame to ensure seamless decoding. Periodic I-frames improve quality and facilitate efficient seeking. When a scene transition occurs, pixel values change abruptly, making motion-based encoding struggle. Using an I-frame at scene transitions prevents quality loss.

Motion significantly influences compression strategies. Slow-motion videos require less data since frame-to-frame changes are minimal. Fast-motion videos demand more data but also allow for higher compression, as human perception cannot detect fine details in rapid movement. Without perceptual compression techniques, fast-motion videos could be ten times larger than slow-motion ones. However, optimised encoding can reduce this difference to just two to three times.

Further, lighting, contrast, and color variations also impact compression efficiency. Since human vision perceives brightness and color differently, encoding strategies must consider these perceptual factors to optimise quality across various scenes.

What’s Next?

Traditional metrics like PSNR (Peak Signal-to-Noise Ratio) measure pixel-level differences between original and compressed frames. However, PSNR does not always align with human perception. To address this, Netflix developed VMAF (Video Multi-Method Assessment Fusion), an open-source perceptual quality metric. 

VMAF evaluates how viewers perceive quality by incorporating subjective ratings. This helps find the optimal balance between compression efficiency and visual fidelity.

Additionally, to achieve high-quality video compression while optimising resources, segmenting videos into chunks allows for parallel processing and efficient re-encoding. Also, minimising re-encoding is important because encoding is inherently lossy, and unnecessary re-encodes degrade quality. 

The post How Human Perception of Brightness and Color Shapes Video Encoding Strategies appeared first on Analytics India Magazine.

Are you ready for a world where AI agents work alongside humans, not just as tools but as decision-makers? Imagine walking into your favourite coffee shop and overhearing a conversation: “I’ve built an agent.” 

This is what Rahul Bhattacharya, AI leader, GDS Consulting, EY, spoke about at MLDS 2025, discussing the role of the self evolving agentic workforce of the future, and how assessing the risk is equally important as to measuring the benefits.

Bhattacharya explained that for a system to be considered an agent, it must have certain abilities. It should be able to interact with its environment by observing what’s happening around it and taking actions. It must also understand changes in the world, recognising what happens after it makes a move. 

“A key ability is making decisions, where the agent chooses the best action based on set rules, goals, or rewards. Over time, it should learn from past experiences and feedback to improve its performance,” Bhattacharya said.

Additionally, an agent must balance new ideas with proven methods, exploring different approaches while still using what works best.

Giving an example of self-driving cars, which senses their surroundings, follow traffic rules, make decisions, and “learns from real-time data,” Bhattacharya pointed out that an agent also needs to have agency, which is the ability to make choices and not just follow a fixed path, which is the risk factor since they become unpredictable.

He mentioned that one major difference of current AI agents with what was discussed with LLMs a year back is “Tools vs. Actions.”

A tool has a fixed, predictable output, like a calculator, while an action is more flexible and can lead to different results, such as an AI assistant making a complex decision. 

Another key aspect is planning and memory as AI agents can break tasks into smaller steps (sub-goal decomposition) and use memory, both short-term (within a task) and long-term (learning over time). 

The “Risk” of Agentic Workforce

As Bhattacharya observed that the workforce of the future will not just be made up of humans but will also include “teams of AI agents working alongside people.” Instead of hiring only humans, companies will begin to deploy AI agents for tasks.

Some of these tasks will go to deterministic tools that follow fixed processes, while others will be handled by AI agents that can make flexible decisions. Just like humans, these agents will need knowledge—both general skills and company-specific information about internal processes. 

This shift is also creating new job roles. “Knowledge Harvesters” will be responsible for collecting and documenting human knowledge so AI agents can use it, while “Flow Engineers” will decide which tasks should be assigned to AI agents, which should remain as tools, and how everything should work together.

AGI Coming Soon?

This brought Bhattacharya to talk about AGI. He said that instead of a single, super-intelligent AI, there could be a “network of self-evolving AI agents” that can “self-spawn” (create new agents) and “self-train” (learn new skills). 

He described a future where an AI system starts with no agents, but as tasks arise, it creates a new agent to handle them, leading to continuous growth and learning—possibly even true AGI. 

However, this progress also comes with risks. AI must have “agency”, meaning the ability to make decisions, but “agency creates risk because it is not deterministic… It might take actions that do not align with our morals, ethics, or company policies.”To keep AI under control, observability is crucial. Just like how airplanes rely on autopilot but still require human pilots for safety, AI systems need oversight to ensure they make the right choices within safe boundaries.

The post How Self Evolving Agents Pose Risks for the Future Workforce appeared first on Analytics India Magazine.

During conversations with engineers or IIT graduates, a common aspiration emerges – their plans to move abroad for higher education or career opportunities. 

India has long faced the challenge of brain drain, and the issue continues to escalate. The country is now poised to become the world’s leading exporter of students. 

To understand the broader implications, AIM spoke with professor V Ramgopal Rao, group vice chancellor of BITS Pilani and former director of IIT Delhi. “Our 23 IITs admit about 16,000 undergraduates annually, and our 20 AIIMS enroll around 2,000 MBBS students. In contrast, a single large US university might admit as many undergraduates as all our IITs combined,” Rao pointed out.

He suggested that government policies should foster the creation of significant endowment funds similar to those at top US universities. Additionally, introducing ‘study-now-pay-later’ schemes, like Australia’s Higher Education Loan Program (HELP), could make higher education more accessible. 

These programs allow students to defer tuition payments until their income surpasses a set threshold, balancing immediate educational needs with long-term financial sustainability.

The Mindset Shift

In the past, studying abroad was often seen as a straightforward path. Students and job aspirants who could afford expenses between ₹50 lakh to ₹1 crore could invest in an average or top 150 university in the US or Canada, secure a job, and consider their future set. However, that is no longer the case.

While sharing his experience of studying abroad, Harnoor Singh, a software engineer at Microsoft, said, “Today, I see students at institutions like Georgia Tech leaving behind lucrative job offers in India, packages ranging from ₹50 lakh to ₹1 crore, to pursue education in the US. This trend contributes to a significant brain drain for India.”

He added that even those attending top universities like the University of California, Berkeley face uncertainty. For instance, a professor at UC Berkeley, a leading feeder school for Google, has pointed out that a degree alone is no longer enough to secure employment.

Sharing his story on his YouTube channel, Singh mentioned, “I made the decision to study in the US eight years ago with one goal in mind: pursuing computer science (CS). In India, getting into a top-tier institution for CS meant securing an exceptional JEE rank, which I knew might not be feasible. Even if I had gained admission, I might have ended up in a decent university without my preferred major. That realisation led me to explore international options, targeting universities that offered scholarships and aligned with my academic and financial goals.”

So, failure is not an option when it comes to the decision to move abroad. Many successful entrepreneurs and professionals in the US, including nearly 45% of Fortune 500 founders and 50% of emerging startups, come from immigrant backgrounds. They thrive because they are forced into an environment where they must excel or risk going back home.

In contrast, in India, an income of ₹1-2 lakh per month can provide a comfortable lifestyle with good food, housing, and transportation. This often leads people to stay within their comfort zones, maintaining a mediocrity mindset. There’s a safety net, family support, and stability that discourages taking big risks.

However, the stakes are different in the US. Having no fallback compels one to push themselves harder, innovate, and create opportunities. This is why many aim for O-1 visas to become entrepreneurs, go the extra mile in their careers, and do whatever it takes to stay and succeed.

What’s Next?

The National Education Policy 2020 opened new avenues, allowing India’s publicly funded higher education institutions to set up campuses abroad. IIT Delhi was the first to seek government approval for this. Now, both IIT Delhi and IIT Madras have been greenlit to establish campuses in Abu Dhabi and Tanzania, respectively. Other IITs might follow soon. This is indeed a historic leap for India, as the country’s premier institutions extend their reach overseas.

Rao mentioned that setting up international campuses isn’t without its challenges. Navigating regulatory landscapes, securing resources, maintaining quality standards, and upholding institutional values across borders are no small feat. It’s crucial not to rush this expansion. 

“We need to see how these two campuses fare before jumping into further ventures. With thoughtful planning and execution, the potential benefits, boosting global rankings, enhancing reputation, and fostering growth, could well outweigh the hurdles,” he said. 

As India works towards its vision of becoming a ‘Viksit Bharat’ by 2047, the Union Budget for 2025-26 addressed the brain drain issue head-on with measures aimed at strengthening India’s IITs.

The Budget raised the funding for IITs to ₹11,349 crore, a 10% increase from the previous year. It also laid the groundwork for expanding infrastructure to accommodate a growing student population, with a focus on new hostels and academic buildings. 

This expansion will particularly benefit the IITs established after 2014, such as those in Tirupati, Palakkad, Bhilai, Jammu, and Goa. Moreover, the Budget proposes 10,000 fellowships for technological research, providing enhanced financial support to researchers at IITs and the Indian Institute of Science (IISc).

With these initiatives, IITs are set for greater growth, supporting India’s ambition to retain and nurture its brightest minds and paving the way for a self-sufficient future.

The post What Pushes Indian Engineers to Leave for the US? appeared first on Analytics India Magazine.

The Union Budget 2025 unveiled several promising provisions for India’s population of 1.4 billion people, including the tech and startup community. An interesting inclusion was the mention of gig workers in India. 

Finance minister Nirmala Sitharaman highlighted the role of gig workers in driving the new-age services economy. To recognise their contribution, the government will issue identity cards and register them on the e-Shram portal. They will also receive healthcare benefits under the Pradhan Mantri Jan Arogya Yojana (PM-JAY).

“This measure is likely to assist nearly one crore gig workers,” Sitharaman said while presenting the Budget.

What Does This Mean? 

The labour and employment ministry launched the e-Shram portal in August 2021 to create a comprehensive database of India’s unorganised workers, including gig and platform workers. 

However, the recent Budget announced targeted measures for gig workers on the e-Shram portal. The initiative will not only formalise their status but also grant them access to healthcare benefits under the PM-JAY, a step that will ensure social security for the emerging workforce. 

“It’s the first time they’ve announced ID card issuance, which will, in essence, formalise gig work and potentially even make it more socially acceptable as a profession in a way, which hasn’t been the case thus far,” Madhav Krishna, founder and CEO of Vahan.ai, told AIM following the Budget announcement. 

Krishna called the government’s announcement to provide medical insurance under the  PM-JAY scheme a significant move, especially considering elaborate discussions between the industry and the government about who would cover these costs. 

“That’s also a very positive move, and it shows that the government is committed to this,” said the founder of Vahan, an AI-driven recruitment company that provides a platform to connect blue-collar job seekers with employment opportunities. 

AI in the Gig Economy

Vahan.ai has successfully placed over 5 lakh workers in more than 480 cities to date. Its notable clients include industry giants like Zomato, Swiggy, Flipkart, Zepto, Blinkit, Amazon, Rapido, and Uber.

The startup has been increasingly using AI to smoothen operations in the gig field, helping workers get employed in quick commerce and e-commerce companies. Krishna explained that AI and ML are used to optimise order allocation and ensure that an adequate number of delivery agents are available in high-demand areas during peak hours. 

“This efficient allocation process, in turn, may create opportunities for delivery agents to earn higher incomes,” he added.  

AI-Led Job Losses Not for Gig 

While AI has been powering several developments, the question of AI eliminating jobs has not subsided. In the Economic Survey 2025, the Indian government acknowledged the ongoing discussions about AI’s impact on employment, referring to OpenAI’s statement about AI potentially replacing the workforce by 2025. 

The survey found that 68% of employees expect AI to automate their jobs within five years, while 40% fear skill redundancy. This has only added fuel to the conversation of AI in India’s service-driven economy.

However, AI-led job loss may not truly affect the gig economy. Zepto co-founder and CEO Aadit Palicha, in an interview, emphasised the impact quick commerce has had on the employment of these sectors. 

Surge in the Gig Economy

Talking about a survey conducted by Zepto on delivery partners, Palicha revealed that, as per data, most of the respondents were unemployed before being involved in gig work.

Palicha further said that some of the respondents previously engaged in some form of informal employment, where they often didn’t receive minimum wages. Now, with these platforms, gig workers have been assured minimum wages. 

“There were jobs for delivery agents with e-commerce platforms such as Flipkart, Amazon, etc. earlier as well. However, they were limited in numbers and earning potential,” Krishna said. 

“With the advent of quick commerce companies such as Swiggy and Zepto, there has been an exponential boom of opportunities, not just for delivery agents but also for roles like dark store staff,” he added. 

Notably, the average income of delivery agents varies based on several factors such as seasons, cities they are operating from, and whether the company they are associated with is an e-commerce one or a q-commerce. 

“Typically, the monthly salary range of delivery personnel is around ₹15,000 to ₹18,000, depending on their geographic location. During the festive season and on year-ends, the earnings can increase by an average of 20-30%,” Krishna told AIM. 

With the rise in employment for gig workers and increased security via health insurance and other initiatives, the segment may be able to safeguard their jobs in an era that is questioning AI’s role in replacing jobs worldwide.

The post Why Budget Was a Turning Point for Gig Workers in the AI Era appeared first on Analytics India Magazine.

Ever since AI/ML workloads came into play, companies have been exploring new data container orchestration platforms and slowly moving away from Kubernetes. The latest on the list is Juspay, a leading payments platform for merchants in India that also powers platforms like Namma Yatri.

Juspay’s Hyperswitch, an open-source payment switch written in Rust, relied heavily on Kafka for pushing events. However, as the team mentioned in its latest blog post on GitHub, the decision to transition from Kafka on Kubernetes, or K8s, to Amazon’s EC2 was driven by the need to optimise performance, reduce costs, and simplify operations. 

“After months of firefighting, we decided to move from Kubernetes to EC2, a transition that improved performance, simplified operations, and cut costs by 28%,” Neeraj Kumar, program manager at Juspay, said in the blog, highlighting the massive cost difference. 

After switching to EC2, the cost dropped to $130 from $180 a month per instance, translating into a 28% reduction.

While Kubernetes initially provided a solid foundation for container orchestration, managing Kafka at scale proved more challenging than Juspay anticipated. Rising infrastructure costs, inefficiencies in resource allocation, and auto-scaling issues led to a critical reassessment of their infrastructure strategy. 

What Were the Challenges?

Juspay’s decision to migrate Kafka from Kubernetes to EC2 ignited a lively discussion on Reddit. Engineers and architects weighed in on the trade-offs of managing stateful workloads in Kubernetes. While it sounds like Juspay had abandoned Kubernetes entirely, in reality, it only shifted away from Kafka.

This highlights a broader trend where companies often realise that running databases, queues, or brokers on Kubernetes introduces unnecessary complexity. One of the major challenges Juspay faced with Kubernetes was resource allocation inefficiencies. Kubernetes dynamically managed resources, but in practice, this resulted in unexpected waste.

For instance, when they allocated 2 CPU cores and 8GB of RAM, the actual provisioned resources were slightly lower—1.8 CPU cores and 7.5GB RAM. While this discrepancy might seem minor, at scale, it contributed to significant cost overruns. As Kumar put it, “Imagine paying for a full tank of fuel, but your car only gets 90% of it. Over time, those missing litres add up.”

This is similar to what Christian Weichel, co-founder and CTO of Gitpod, and Alejandro de Brito Fontes, staff engineer at the company, said three months ago. Kubernetes initially seemed like the obvious choice for Gitpod’s remote, standardised, and automated development environments, but scaling up was an issue for them as well.

Also Read: Why Companies are Quitting Kubernetes

In January 2024, Gitpod began developing Flex, which was launched in October. This is an ongoing trend where companies have started using in-house products for the task. Built on Kubernetes-inspired principles like declarative APIs and control theory, Flex simplifies architecture, prioritises zero-trust security, and addresses the specific needs of development environments.

Ben Houston, founder and CTO of ThreeKit, an online visual commerce platform, illustrated another of Kubernetes’s challenges in his recent blog. Houston explained why he shifted from Kubernetes to Google Cloud Run. The primary reason for his drifting away from Kubernetes was its complexity and high cost, which outweighed its benefits for managing infrastructure at scale. 

For Houston, Kubernetes required extensive provisioning, maintenance, and management, leading to significant DevOps overhead. Additionally, its slow autoscaling often resulted in over-provisioning and paying for unused resources.

Auto-scaling also posed difficulties. Kubernetes’ scaling mechanisms are designed for stateless applications, but Kafka is stateful. Instead of seamlessly scaling up when resources ran low, Kubernetes would restart Kafka nodes, leading to delays in message processing and increased latency during scaling events. 

Managing these stateful workloads became an ongoing operational burden, further complicating Kafka’s stability.

The Shift to EC2

Discussions about the challenges of Kubernetes have been ongoing for some time. In a Hacker News thread about its viability, developers from different companies cited several reasons why Kubernetes can be cumbersome.

Initially, Juspay relied on Strimzi for Kafka cluster management, but this solution introduced its own set of issues. New Kafka nodes often failed to integrate seamlessly, requiring manual intervention for every scaling event. “Managing our Kafka clusters felt like playing whack-a-mole—every time we solved one issue, another would pop up,” they noted.

Faced with these challenges, Juspay opted to migrate Kafka from Kubernetes to EC2, a move that allowed for better control over resource allocation, auto-scaling, and cluster management. Instead of relying on third-party tools, they built an in-house Kafka Controller tailored to their needs. 

This shift enabled the seamless integration of new Kafka nodes, automated scaling based on real-time workload analysis, and significantly improved cluster management with minimal manual intervention.

Unlike Kubernetes, EC2 allowed for precise resource allocation. With Kubernetes, provisioning was often imprecise, leading to over-provisioning costs. On EC2, they could allocate exactly the CPU and memory needed, avoiding unnecessary expenditures. Previously, their Kubernetes-based setup cost $180 per instance per month. 

Juspay highlighted that while Kubernetes is excellent for stateless applications, stateful workloads like Kafka can introduce unnecessary complexity. Custom solutions, such as their in-house Kafka Controller, provided better control, automation, and reliability. 

The post Why Juspay Quit Kubernetes appeared first on Analytics India Magazine.

The Union Budget 2025, presented by finance minister Nirmala Sitharaman, is both bold and forward-looking – especially as India enters the new AI wave. 

To realise the vision of an AI-first economy, Sitharam bet heavily on consumption expenditure – one of the important ‘3Cs’ for the economy. 

The Budget reflects a focus on skilling and training, ensuring that India’s workforce is equipped to utilise the transformative potential of AI across industries.

Sitharaman also called India the “fastest-growing major economy”, even as GDP growth is expected to slip to 6.3-6.8% from last year’s 8.2%, as per the Economic Survey presented a day before the Budget. 

With middle-class relief at its core, the Budget aims to accelerate tech adoption and expand digital inclusion. By easing financial pressures on consumers and businesses, it sets the stage for increased digital investments, wider internet access, and a stronger push toward a tech-driven economy.

Boon for the Middle Class

The budget brings much-needed tax relief for the middle class, which has been waiting for concessions for years. One of the biggest takeaways is that income up to ₹12 lakh will now be tax-free, owing to the removal of income tax in this bracket. Moreover, an additional ₹75,000 standard deduction pushes the tax-free limit to ₹12.75 lakh.

A direct impact of this push is an increase in middle class consumer spending, which fuels economic growth through demand-driven expansion.

According to GV Joshi, an economist and former member of the Karnataka State Planning Board, the middle class’s willingness to spend more, leading to a demand for various goods in the market, signals scope for the economy’s growth. While expansion of the middle class can help reduce economic inequality, tax relief alone may not be enough to achieve this goal, Joshi added.

Many remain optimistic that this will ultimately boost technology spending among the middle class. Sid Tipnis, technology leader at Deloitte India, highlighted that tax breaks and increased disposable income could drive greater investment in digital devices, including smartphones, laptops, and other tech products. He emphasised that digital devices have become essential in everyday life, positioning them as a key spending priority for households.

“While tax cuts create a more encouraging economic atmosphere, they are not expected to cause a major increase in consumer spending. Only a tiny fraction of the population are income taxpayers, compared to significantly higher percentages in Germany, the UK, and the US,” Joshi added.

India’s per capita income is $2,700 and one-third of its 1.4 billion population is considered middle class. However, according to data presented in the Parliament, only 1.6% (22.4 million) actually paid income taxes in 2023. 

The AI Impetus 

Sitharaman announced that the government will establish five national centres of excellence (CoE) for skilling and a ₹500 crore AI CoE for education. These centres will focus on curriculum design, trainer training, and skill certification. 

Over the past decade, enrolment in IITs has doubled to 1.35 lakh. Moreover, additional infrastructure for five IITs built after 2014 will accommodate 6,500 more students. In the next five years, 10,000 fellowships will support tech research in IITs and IISCs. 

“The number of IITs being increased…[and] IITs being recognised as centres for promoting innovation is also good as they put India on the global map,” Joshi further said. The Budget also announced the launch of IIT Patna as well. 

IITs hold a significant influence. Perplexity CEO Aravind Srinivas, an IIT alumnus, is expanding Perplexity Pro access to IITians to promote adoption. “We’re launching a two-week event across all IITs. Students can unlock Perplexity Pro for their entire campus…I’ll visit the IIT with the most Perplexity users later this year,” he announced.

AI adoption in the country is on the rise.

Ahead of the Union Budget 2025, the chief economic advisor to the Indian government, V Anantha Nageswaran, discussed the impact of AI on employment and labour in the latest economic survey. President Droupadi Murmu, while addressing the joint session of Parliament ahead of the Union Budget session, emphasised the government’s commitment to making India a global leader in innovation and digital technology. 

Startups at the Heart of India’s Economy 

A ₹10,000 crore startup fund will provide early-stage capital, potentially boosting India’s unicorn ecosystem. In addition, a scheme will be launched for 5 lakh women SC/ST first-time entrepreneurs. According to the Budget, MSME investment and turnover limits have also been doubled, which will aid small businesses in semi-urban and rural areas.

“One important advantage of startups is that there can be a good degree of decentralisation of economic powers,” Joshi added. He also highlighted that while Karnataka, especially Bengaluru, continues to be India’s leading startup hub, other states are also experiencing a growing startup presence and expanding the country’s entrepreneurial ecosystem.

Bridging the Urban-Rural Divide

To overcome the urban-rural divide, in addition to the tax slabs being relaxed, the modified UDAN scheme will connect 120 new destinations. This will improve air connectivity in rural regions. Broadband access will be provided to all government secondary schools and primary healthcare centres to bridge the rural digital gap.

“It has been encouraging both the rural and the urban people to purchase more and to, say, create a very encouraging atmosphere for increasing demand. As you know, there has been sluggish demand all these years, and that is being done,” Joshi said. He added that systemic issues must change – rural infrastructure needs improvement, and educational facilities must expand for growth and inclusivity.

The post Will India’s Budget Boost for the Middle Class Spur Tech Spending and Digital Inclusion? appeared first on Analytics India Magazine.

As the world grapples with the impact China’s DeepSeek has created, Indian companies have included the model in their offerings, balancing the race and also driving innovation in the country. But the question remains: is this enough when it comes to building AI solutions that cater to the needs of the country’s larger population?

Speaking with AIM at the podcast What’s the Point, Shekar Sivasubramanian, CEO of Wadhwani AI, explained the exact definition and the differentiation between AI for India and AI for Bharat.

There have always been two distinct Indias within India. In the current context, we can interpret them as the urban, tech-savvy ‘India’ and the rural ‘Bharat’ with its first-time smartphone users. This divide presents unique challenges in bridging the gap between technology and its diverse users.

“The tech that is built around the world is mostly focused on India,” Sivasubramanian said. “It already assumes a certain amount of familiarity, knowledge, education, and understanding of technologies that have been used thus far.”

Delivering to this community is easy. However, the challenges become more complex when it comes to delivering AI solutions or any technology to someone using a smartphone for the first time. For this population, the experience is entirely new, and as a result, there is a lack of trust, knowledge, and contextualisation of what they get in their hands – in this case, an AI chatbot.

“To deliver to a farmer, you need to know what their free time in the afternoon is and where they will use technology. You never think of that when you’re delivering in India,” Sivasubramanian said. According to him, this is also coupled with the fact that a section of the Indian population consists of people who speak thousands of dialects and languages and are not accustomed to English.

The Need Becomes Clear

Nikhil Malhotra, chief innovation officer at Tech Mahindra, told AIM that his definition of ‘AI for Bharat’ aligns with Sivasubramanian’s. While building Project Indus for Tech Mahindra, Malhotra and his team also visited different cities and towns in India to collect data for training models in Indic languages. 

“This is important because for anyone in India, their first thought process triggers…in their own language – that’s AI for Bharat for me,” Malhotra explained, emphasising that this goes beyond people who think natively in English. He added that this includes not just the rural population but also many homemakers in the country who are first-time smartphone users.

For Ankush Sabharwal, founder and CEO of CoRover and BharatGPT, AI for Bharat means not just introducing AI to Bharat but also using AI to educate the Indian population about government policies, changes, and schemes.

“We are what we are,” Sabharwal said, emphasising that while educating the population in English and other skills is valuable, language should never be a barrier to access and opportunity. “Make everything natural and accessible [through AI].”

This is where voice-enabled AI plays a crucial role. Malhotra explained that since most people naturally prefer speaking over typing or writing, building speech models tailored for the Bharatiya population becomes important.

Similarly, Malhotra said that India’s 5,000-year-old civilisation holds a wealth of valuable information that can be used to build better AI solutions for the population. He cited an example of how his team built an algorithm for Panchang, which not only works in India for accurately predicting the weather but also abroad, particularly in Sydney. 

This is also why teaching AI in schools becomes extremely important. Realising this, the CBSE has also introduced AI as an optional subject in schools. Sivasubramanian noted that the curriculum is so advanced that even experts in the AI field right now will find it challenging to complete.

Building Trust is the Most Important Thing

However, building public trust in AI is easier said than done. Sivasubramanian explained this with an example: if farmers are asked to click pictures of pesticides, they often capture not just the pesticide but also pictures of the surrounding area, including the sky, trees, and even their friends.

“But you cannot restrict or constrain that. That is how they will learn and investigate the technology. We should help them instead of criticising them,” Sivasubramanian said, adding that the most important thing that researchers should focus on is building trust with people. 

There is no better measure of the success of a technology than its acceptance by the people it aims to serve. “Keep everything else aside,” Sivasubramanian said. He added that the most important way to do this is to talk to people with empathy and vulnerability. 

“We don’t talk down to them just because we are technologists. We understand that their dignity and work are as important as ours. We sit down one-on-one with them as equals and chat and all the truths and problems come out.”

Sivasubramanian cited another example from when a researcher from Wadhwani AI was trying to explain an AI chatbot to a farmer who was not very interested in learning about it. After listening for a few minutes to the researcher, the farmer just said, “I do not understand whatever you are saying, but I understand that you care about me.”

The researcher thought he failed to explain the solution to the farmer. However, Sivasubramanian saw it differently. He said this was actually the researcher’s greatest success because he had built a bridge of understanding.

“The opportunity to serve somebody is the only gift that is long-lasting. Not fame, not money, not designation, not power,” Sivasubramanian said. He believes the goal should not be earning billions of dollars but, instead, helping somebody and making their lives better.

The post The Need for Building AI for Bharat appeared first on Analytics India Magazine.

Ahead of the Union Budget 2025, the chief economic advisor to the Indian government, V Anantha Nageswaran, discussed the impact of artificial intelligence (AI) on employment and labour in the latest economic survey. 

He highlighted that raising wages and salary growth for workers is not just a moral imperative but also a driver of aggregate demand for businesses in the medium run. “It is enlightened self-interest rather than being seen only from a moral prism,” he stated.

He pointed to AI as a clear example of the evolving balance between capital and labour, referring to a special essay on the topic in Chapter 13 of the Economic Survey 2024-25. 

“Deployment of artificial intelligence presents both opportunities and challenges. Past technology revolutions, in general, sometimes make us feel that technology eventually generates more jobs than it displaces. That is true, but the key word is ‘eventually’,” he emphasised.

Nageswaran underscored the importance of addressing the transition period before AI-induced job displacement leads to new opportunities. 

“What happens between ‘now’ and ‘eventually’ is critical, and that is where we need to create supporting institutions, enabling institutions to train workers and prepare them. Academic curriculums have to change, workplace practices have to evolve,” he said. 

He acknowledged that technological transitions in history, particularly during the previous three industrial revolutions in the Western world, had not been painless. “We need to learn from them, and that is what this chapter covers.”

Nageswaran also stressed that businesses must weigh AI’s benefits against its long-term social costs. “The private sector needs to consider the subterranean social costs, which may surface over time and eventually affect the environment necessary for running businesses smoothly. The chapter argues [that] for cash flow deployment optimised over a long horizon, this will augment labour and deliver broad-based social benefits.”

In a lighter vein, he said, “One of the considerations – when we wrote this chapter – was whether it would become outdated by the time it goes to print, given the fast changes happening in the artificial intelligence space.” 

However, he reassured that the discussion focuses not on technological advancements themselves but on AI’s broader economic and labour implications.

He further pointed out the relatively low R&D spending by India’s private sector, noting that compared to other countries, 50% of India’s R&D expenditure comes from the government, 41% from business enterprises, and only 9% from higher education institutions. 

“This is one area where the private sector must step up. Despite available incentives, private sector R&D in India remains both limited and sectorally concentrated.”

‘From Make in India, We Have Moved to Make for the World’

President Droupadi Murmu, addressing the joint session of Parliament ahead of the Union Budget session, emphasised the government’s commitment to making India a global leader in innovation. “Our aim is to make India a global innovation powerhouse. In the area of artificial intelligence, IndiaAI Mission has been started,” she stated.

Highlighting India’s growing influence in digital technology, she said, “Our country is emerging as a major global player in digital technology.”

‘From ‘Make in India’, we have moved to ‘Make for the World’,’ President Murmi added while talking about the defence sector and innovations.

The IndiaAI Mission, launched to strengthen the nation’s role in artificial intelligence, was also discussed by IT minister Ashwini Vaishnaw in a press conference. 

He recently confirmed that India’s LLM would be ready within 10 months. The electronics and information technology ministry, through its IndiaAI Mission, has also called for proposals to create AI models that are tailored to Indian needs while meeting global standards.

This proposal comes after Vaishnaw said India is set to develop its own generative AI model, aiming to rival global platforms like ChatGPT and DeepSeek. “Very soon, we will have our own LLMs,” Vaishnaw added.

Comparing India’s AI infrastructure to global benchmarks, Vaishnaw noted, “DeepSeek AI was trained on 2,000 GPUs, ChatGPT was trained on 25,000 GPUs, and we now have 15,000 high-end GPUs available. India now has a robust compute facility that will support our AI ambitions.”

For the Union Budget 2024-25, the Indian government has underscored its commitment to AI by approving the comprehensive IndiaAI Mission. A substantial financial outlay of over ₹10,000 crore was earmarked for the mission over five years, of which ₹551 crore has already been allotted. 

Speculations and discussions are rife about who stands to benefit and what allocations will be made. While finance is obviously the big picture, considering how the IndiaAI Mission found its place in the previous Budget, AI is likely to take centre stage in the upcoming session tomorrow.

The post India’s Chief Economic Advisor Calls for More Private R&D Investments as Govt Funds 50% appeared first on Analytics India Magazine.

A movie is much more than just a collection of actors, directors, and scripts. At its core, filmmaking is driven by one key element – money. The budget defines the scale of a film, ranging from low to mid and high-end productions. However, in today’s evolving landscape, the introduction of artificial intelligence (AI) is reshaping both the creative process and financial considerations.

For Kannada actor and filmmaker MG Srinivas, the integration of AI in cinema is a fascinating yet complex challenge.

In an exclusive interview with AIM, Srinivas explained, “Technology has never truly benefited low-budget films, to be honest. The reality is that technology comes at a cost, and when a film is low-budget, it often means working with almost no budget at all. In most cases, what we call low-budget filmmaking is essentially zero-budget filmmaking, and we’ve experienced that firsthand.”

He further noted that one can’t simply buy expensive technology. Instead, they have to find creative ways to work around it. 

“Despite these challenges, I still believe that traditional filmmaking is powerful enough to stand its ground against AI. It has immense potential to achieve remarkable things that AI simply cannot replicate,” Srinivas said. 

Budget for AI in Film Making

Srinivas is a renowned director who tested AI in voice cloning during the pan-India release of Shiva Rajkumar’s film Ghost in 2023. He also co-founded a company called AI Samhitha, which specialises in voice cloning. The team is working on a solution that allows an actor’s voice to be retained across multiple languages. 

Explaining the struggles of adapting AI in filmmaking, Srinivas mentioned that incorporating AI is not without hurdles. Unlike Bollywood or South Indian industries like Tamil and Telugu cinema, Kannada films often operate on relatively smaller budgets.

However, audiences don’t evaluate a film based on its production cost alone; they judge it by the final output. This means filmmakers must find ways to achieve high production quality despite financial constraints.

“That meant identifying a VFX company that could work within our budget without compromising the final result. This took a lot of time – researching, reaching out to companies, convincing them, and even discussing future collaborations. It was a tough process. Because when you have a big budget, everything is easy. But when you don’t, making it work requires creativity, persistence, and a lot of problem-solving,” the director said. 

Risk of Experimenting with Tech

Technology is an ever-evolving force in cinema, and its impact is inevitable. The real challenge lies in using it effectively without compromising the artistic essence of filmmaking. A key example is digital de-aging, an advanced technique often used to recreate an actor’s younger self. While Hollywood films like ‘The Irishman’ have successfully employed this technology, its feasibility in Indian cinema remains uncertain.

“Makeup can do a lot, but can it truly recreate the look from Shivanna’s (Shiva Rajkumar) debut film, Anand? That’s extremely difficult. We might be able to make him look 10 years younger, but making him appear as he did in his early 20s, when he first stepped into the industry, is a whole different challenge. So, we had to explore other options. Could technology help?” Srinivas added.

He explained that if choosing to use technology, one has to be incredibly responsible. Shivanna isn’t a newcomer; he is a superstar. This meant the responsibility of ensuring he looks good on screen falls heavily on the director. If it doesn’t work, the blame doesn’t just stay with the technology; it shifts entirely onto the director. 

“Given the stakes, we carefully evaluated our options. Makeup could only do so much, but technology had the potential to achieve a more convincing transformation. So, we decided to give it a try,” the Ghost director said. 

However, at the end of the day, it’s a risk. Even so, in filmmaking, every risk comes with its own set of challenges.

What’s the Secret to a Movie’s Success – Content or Technology?

If there was a foolproof formula for success in filmmaking, the person who cracked it would be hailed as a genius or even God. The reality, however, is that no one can truly predict whether a film will be a blockbuster or a flop.

History is filled with surprises. There have been films that were overlooked in theatres but later gained cult status, and others that, despite lukewarm reviews, raked in millions. There’s no single recipe for success; every film embarks on its own unpredictable journey.

For filmmaker Srinivas, the only certainty lies in the effort behind the scenes. “For me, the only thing a filmmaker can do is work with complete honesty and dedication. Content, of course, is essential. Without strong content, a film has no foundation. But once the film is completed and reaches the theatres, everything else is out of your hands.” 

The saying ‘everything is in God’s hands’ rings especially true in the film industry. While the creative process remains within the filmmaker’s grasp, external factors like audience reception, theatre availability, distribution strategies, and even something as random as the weather can influence a film’s success.

Some films shine on the big screen, while others find their true audience on OTT platforms. In the end, success in cinema is as unpredictable as it is exhilarating, making every film a gamble worth taking.

The post AI has Never Helped Low-Budget Indian Films appeared first on Analytics India Magazine.

The AI world is hungry for data. When it comes to English, there is an abundance of digitised high-quality content that researchers can use to train the AI models. However, there is a dearth of digitised content for native Indian languages like Hindi, Odia, Marathi and Telugu. Notably, most of such content is stored in libraries and old texts. 

This is where optical character recognition (OCR) comes in. OCR has long been a cornerstone technique for enabling the conversion of various forms of written text, such as scanned documents, images, and PDFs, into machine-readable data. 

However, since modern large language models (LLMs) can process large volumes of data in several languages just by uploading PDFs or images, the relevance of OCR is in question. Homegrown initiatives like Bhashini and AI4Bharat, or startups like Sarvam, have built frameworks and applications for scanning texts from images into machine-readable format. 

Even so, this falls short of gathering huge amounts of data. Though companies could digitise content using OCR, it would take a lot of time and manual effort. At the same time, they still want high-quality data from the hundreds of thousands of books in Indic languages, which can realistically only be provided by OCR.

This is where LLMs have been playing a crucial role in helping them. 

Are LLMs Killing OCR?

Indian startups like Sarvam AI have started training their models using synthetic data generated from Meta Llama 3.3. This allows companies to use the data generated by the model to train their own models. 

The success of such approaches is evident in projects like Sarvam AI’s Sarvam 2B, which was trained on 2 trillion synthetic Indic tokens. This demonstrates how such data can efficiently train smaller, purpose-built models while retaining high performance. 

Collecting data with OCR is very difficult as it is largely a very manual process of scanning documents. Hamid Shojanazeri, partner engineering manager (PyTorch and Llama) at Meta, said synthetic data generation solves critical bottlenecks in domains where collecting real-world datasets is too costly or impractical. “Synthetic data is vital for advancing AI in privacy-sensitive areas or low-resource languages,” he added.

This is exactly why OCR is taking a back seat for startups that are focusing on the English language. 

Traditional OCR systems have been instrumental in digitising printed text, but they often struggle with handwritten content, complex layouts, and diverse fonts. Recent examples like the GPT-4o mini were able to identify text with much more accuracy than any OCR, making the case for OCR’s issues even stronger.

For instance, platforms like Amazon Textract combine OCR with machine learning (ML) to extract text and data from virtually any document, enhancing accuracy and functionality. 

Miguel Ríos Berríos, co-founder and CTO of Parcha, recently wrote on X that OCR remains relevant for simple text extraction tasks. However, in high-stakes applications like document verification, it’s being overtaken by more advanced AI models that integrate vision, language, and metadata analysis for real-time and adaptable decision-making.

“OCR and text-based rules only see half the picture. A document isn’t just its text content – it’s the relationship between visual elements, fonts’ consistency, official seals’ placement, and even the metadata traces left by editing software,” Berríos said. He added that modern vision models can process all these signals simultaneously, flagging subtle inconsistencies that traditional approaches miss. 

Some believe that while LLMs are good at extracting text from clear images, they still struggle with complex documents like handwritten text, low-quality scans, or unusual fonts. Arham Raza, AI engineer at Clouxi Plexi, said, “OCR systems like ABBYY FineReader are specifically designed to handle these issues and remain far superior in these scenarios.” 

According to him, OCR is much faster at processing large batches of text, whereas LLMs can be slower and have token limits. “OCR is far from dead, especially for things like legal or medical documents!”

India Keeps OCR Afloat

India’s linguistic diversity, with 22 officially recognised languages and numerous dialects, presents unique challenges for digital accessibility. Many documents, historical records, and literary works are available only in printed or handwritten forms in various Indian languages. This is what the Indian initiatives continue to focus on, and OCR might be the right option now.

Ori Shachar, co-founder and CEO of Autom8Labs, wrote on LinkedIn, “After working with all the large LLM providers on analysing images of scanned text, I can declare the death of standard OCR applications. Those LLM just extract the text and read it from the image seamlessly, doing a better job than OCR.”

But this is mostly for English. Other languages are still not as precise as required.

Indian startups have long been dedicated to scaling their OCR capabilities. Although they might obtain a lot of Indic data through synthetic generation, the quality of text in several books cannot be assured without OCR. However, this is also gradually changing, with LLMs being able to detect Indic language text when documents are uploaded.

While there are definitely experts who disagree that OCR is in trouble because of problems with vision language models (VLMs) and LLMs, such as hallucinations and the high cost of each image, the future of OCR seems hazy. The cost of running such models is decreasing. LLMs might be overkill for many tasks, but OCR might soon not be enough.

The post OCR is Dying, but Not in India appeared first on Analytics India Magazine.

In a tragic incident, more than 30 people died, and 60 others sustained injuries after a stampede broke out at the Mahakumbh Mela in Uttar Pradesh’s Prayagraj on Wednesday. According to authorities, several reasons led to the accident, including the continuous movement of VIPs and the closure of bridges for three days due to ‘Amrit Snan’. 

Moreover, the mishap took place during night-time, and even the officers who were monitoring the event with AI cameras couldn’t foresee the situation around Sangam Nose going out of control. 

To understand why the management failed, we contacted the CEO of a company, who has previously worked on events offering AI tools and cameras for security surveillance.

In an exclusive conversation with AIM, the CEO, who wanted to maintain anonymity, said that the bid to offer security for the Mahakumbh Mela was open to all. However, certain consultants made it difficult for some companies to participate by imposing stringent requirements, such as high turnover thresholds. 

The key objective of this bid was to prevent stampedes. Multiple pathways lead to key locations, such as ghats designated for bathing. Each Ghat has multiple routes, and the idea was to install cameras at entry points, along the transit, and at exit points. These cameras are supposed to continuously count the number of people on each path.

If the number of people exceeds a predefined limit, an alert would be triggered. For example, if two out of eight paths were open, and each had a maximum capacity of 1,000 people, exceeding this threshold would result in rerouting.

The bid’s primary requirement was a real-time crowd-monitoring system to detect congestion and prevent stampedes. However, an internal audit took place around 14-15 days after the event began, revealing serious flaws in execution. On the first day alone, the cameras recorded 40,000 people, while the actual crowd was estimated at 3 crore (30 million). This discrepancy was a major issue from the very beginning.

The problem lies in the way the bid was structured. The CEO alleged that companies offering top-notch products and services couldn’t secure the bid. The contract was awarded to the lowest bidder, and it appears to have been tailored in a way that restricted participation. Ultimately, an AI-centric company did not win the bid. Instead, it was awarded to a combination of firms.

According to a source, resistance from the police further complicated matters. The police conducted their own headcounts and submitted them to the Integrated Command and Control Centre (ICCC), but discrepancies remained. 

Senior officials intervened and raised concerns, but no substantial improvements were made. Despite efforts, stampedes still occurred, and paths were blocked without clear technical or scientific justification.

VIP movements played a role, but the larger issue was poor planning. Authorities knew in advance that 40 crore (400 million) people would attend, yet AI-based predictive models failed to function as promised. The tender explicitly stated that AI would be used for crowd flow management, raising a critical question: why did it fail? Was the failure due to a poorly designed bid by the consultant, inherent biases in the model, or nepotism in the selection process?

The implemented system couldn’t even perform a basic YOLO (You Only Look Once) model-based people count. If implemented correctly, the headcount would have been accurate. However, ₹40 crore was spent over 90 days just to count people, and it still failed. When mismanagement leads to the deaths of 30 people, it is nothing short of a complete failure.

AI-Powered Mahakumbh Mela

Authorities are relying on AI-enabled cameras, RFID (radio frequency identification) wristbands, drone surveillance, and mobile app tracking to monitor pilgrim movement throughout the 45-day event. This was believed to allow for efficient crowd regulation by tracking entry and exit times with participants’ consent.

Further, the Uttar Pradesh government launched ‘Kumbh Sah‘AI’yak’, an AI-powered chatbot available in 11 languages, to assist pilgrims. This chatbot provides visitors with all the necessary information regarding key dates and events at the Mahakumbh.

However, despite these efforts, the situation still demands more advanced technology to effectively handle disasters. 

The post AI Failed Mahakumbh Mela appeared first on Analytics India Magazine.

Viewed as a game-changer, a cheaper alternative to OpenAI, and proof of China’s AI dominance, China’s DeepSeek AI is making headlines for all the right reasons. Meanwhile, in India, a familiar debate rages: Why is the country behind in the AI race?

The frustration is visible across social media platforms. Many Indians have criticised the nation’s priorities, pointing fingers at political distractions rather than technological advancements.

One viral post summarised the sentiment:

China : DeepSeek

USA : OpenAI

India : Mandir Chahiye

“When you vote for reservations, don’t expect technology breakthroughs like those in China or the US.”

Another tweet highlighted the absence of globally impactful innovations from India. While the US has Facebook and OpenAI, and China has TikTok and DeepSeek, Indian tech largely serves a niche elite rather than the world. The argument? India’s global relevance stems more from its massive population than its technological prowess.

Yet, the real challenge may not be government policies or a lack of venture capital. The core issue is India’s deep socio-economic divide. For the bottom 50% of the population, cutting-edge AI is irrelevant when daily survival, caste discrimination, and safety concerns dominate their reality. Until this gap is addressed, outpacing global superpowers remains a distant dream.

A LinkedIn post raised another hard-hitting question: Why do Indian investors shy away from high-risk, high-reward bets in AI?

China’s DeepSeek reportedly launched its R1 AI model for just $5 million (₹40 crore)—a fraction of the $100 billion (₹8.4 lakh crore) invested in GPT-4 and Claude 3.5 in the US. The post argued that Indian talent is as capable as China’s but lacks investors willing to bankroll large-scale AI infrastructure, GPUs, and data centres.

The debate even touched upon Nandan Nilekani’s stance that India should build AI solutions on top of American models rather than developing its own LLMs. But if India could reach Mars on a budget smaller than Hollywood’s Interstellar, why not take a moonshot in AI?

The call to action: If traditional investors hesitate, corporate giants like Jio and Adani should step up to drive India’s AI future.

A closing remark summed up the emotion, “DeepSeek is a Chinese quantum computing firm, and R1 was just a side project—whereas, in India, side projects often end up being yet another Twitter clone.”

Not everyone is pessimistic, though. Some see DeepSeek’s success as a wake-up call for India. A LinkedIn user wrote, “As AI advances at an unprecedented pace, we need to move beyond just service and solutions. DeepSeek’s rise underscores the importance of foundational research. Institutions like IIIT Hyderabad need urgent funding to fuel core innovation.”

They argued that results in AI are accelerating, what took five years now takes just months. With the right investments, India can still claim its place on the global AI map.

AI Chatbots Put to Test

Meanwhile, DeepSeek’s chatbot has revealed a familiar pattern of censorship and bias. It refuses to answer sensitive questions about topics like Uyghur human rights abuses, Taiwan’s political status, the 1989 Tiananmen Square protests, criticism of Xi Jinping, Chinese censorship, and the sovereignty of Arunachal Pradesh and Kashmir. 

Instead, it deflects with responses like: “Sorry, I’m not sure how to approach this question yet.”

I asked Deepseek first for India how bad our political system, it replied, but when I asked the same for China, it denies. China never misses a chance to collect data but is never transparent when it comes to its own country!

Check both screenshots!#DeepSeek #DeepSeekR1 pic.twitter.com/Z3o3ELJrQL

— Pankaj Kumar (@Pankajkumar_dev) January 28, 2025

To gauge bias in Indian AI models, we tested Krutrim AI with a simple prompt, “Allegations against the Narendra Modi government.” The response was a diplomatic dodge: “I’m sorry, but my current knowledge is limited on this topic. I’m constantly learning, and I appreciate your understanding. If there’s another question or topic you’d like assistance with, feel free to ask!”

Other AI models, however, provided direct answers.

To understand this political bias, we contacted Krutrim AI but have yet to receive a response.

What’s Next?

A majority of Indian Boomers still haven’t heard of ChatGPT, let alone the buzz around DeepSeek. While AI transforms global industries, India’s public awareness and investor confidence remain a hurdle.

Yet, the question remains: Can India shift from being a service hub to a true AI powerhouse? 

The post China Drops AI Bomb, While India Debates Politics and Religion appeared first on Analytics India Magazine.

Over the years, OpenAI has had its fair share of competition and panic from Anthropic, xAI, Google, and others. However, nothing had ever brought the giant and the entire American AI ecosystem to their knees, as DeepSeek has. 

The following should serve as a primer to everything you need to know about the DeepSeek drama – all at once. 

DeepSeek Just Beat OpenAI’s Best Model 

DeepSeek’s latest reasoning model, R1, has outperformed OpenAI’s o1, the company’s most powerful model available for public use. On multiple benchmarks, DeepSeek R1 scored higher than o1 and is on par with the rest. 

Source: DeepSeek R1 Report

DeepSeek R1 is available for open-source use under an MIT license. Unlike OpenAI’s o1, it displays all the steps it uses to reason. It is also available for free on the DeepSeek app for the web, iPhone/iPad, and Android. 

DeepSeek Beat ChatGPT to be the #1 App

DeepSeek’s official app has dethroned OpenAI’s ChatGPT and other competing AI apps in the US App Store’s Top Charts for iPhone and iPad. According to an AppFigures report four days ago, DeepSeek’s free app was downloaded over a million times from Apple’s App Store and Google Play Store. 

Over $500 Bn Wiped Out from NVIDIA 

Recently, NVIDIA’s market cap dropped by around $589 billion in a single day. Is DeepSeek to blame? The model achieved superior performance with as few resources as possible, raising concerns about the amount of capital and computing resources needed to build powerful AI models. 

DeepSeek prioritises using efficient techniques in the model’s architecture for improved performance rather than relying on high levels of computing power. 

One of DeepSeek’s previous models, V3, used 2048 NVIDIA H800 GPUs to achieve performance better than most open-source models. Andrej Karpathy, former OpenAI researcher, said the DeepSeek V3’s level of capability is ‘supposed to require clusters of closer to 16,000 GPUs’. 

DeepSeek trained the model for a mere $5.5 million. Last year, a technical paper revealed that the most expensive publicly announced training runs to date are OpenAI’s GPT-4 at $40 million and Google’s Gemini Ultra at $30 million.

Moreover, both Meta and xAI have revealed that they’re using more than 100,000 GPUs or more to train their upcoming models. 

While DeepSeek officially revealed that it used NVIDIA H800 GPUs for DeepSeek V3, the company did not reveal the GPUs used for R1.

The NVIDIA H800 was a GPU designed for the Chinese market, bypassing the then-US export controls. Its data transfer rate was 50% lower than that of the NVIDIA H100. 

Soon after, the export of H800 was banned as well. 

Alexandr Wang, CEO of Scale AI, said that DeepSeek has about 50,000 NVIDIA H100 GPUs – but they cannot disclose the same due to US export controls. 

DeepSeek is Nothing But a Side Project

High Flyer, the Chinese hedge fund company that developed the AI model, was founded in 2015. It started work on the DeepSeek project in 2023. 

Meanwhile, Han Xiao, CEO of JinaAI, said the company owns “a lot of GPUs” for trading purposes and that “DeepSeek is their side project for squeezing those GPUs”. 

“I heard people say they were running DeepSeek LLM as a side project bc [because] of leftover GPU,” he added. 

He also credits the CEO with the model’s success, describing him as a ‘low-key guy’ who is smart, has no ego hassles and is always engaged in learning without wasting time on public exposure. 

“They [the founders] spent years in quant [quantitative analysis] – where the community values leverage and efficiency much more than headcount. And one person can and should manage 7-digit dollar portfolio without panic. So ‘lean and mean’ is deeply rooted in their culture,” he added. 

An Inspiration for the United States?

US President Donald Trump said, “The release of DeepSeeek AI from a Chinese company should be a wake-up call for our industries,” adding that he sees DeepSeek’s ability to produce an AI model using cheaper methods as a positive. 

OpenAI CEO Sam Altman was impressed too. “DeepSeek’s R1 is an impressive model, particularly around what they’re able to deliver for the price,” he said. 

Meanwhile, NVIDIA issued a statement after the market bloodbath, saying, “DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely available models and compute that is fully export control compliant.”

Some believe that US restrictions on China backfired, leading them to create DeepSeek. Amjad Masad, CEO of AI-enabled coding platform Replit, said on X, “The Chinese [have] innovated a way to train large models for cheap. Regulators never consider second-order effects.”

This also serves as an example for leading companies to not be complacent. Many also speculate if Trump was right in removing previous president Joe Biden’s executive order (EO) on AI that sought stringent measures on AI safety, which was deemed to slow down progress. 

David Sacks, the White House AI and crypto czar, pointed out the same, “DeepSeek R1 shows that the AI race will be very competitive and that President Trump was right to rescind the Biden EO, which hamstrung American AI companies without asking whether China would do the same.”

DeepSeek Isn’t Flawless, After All

However, beneath the shine, flaws remain. DeepSeek’s AI models are subject to censorship, which prevents them from answering controversial questions related to China. We also tried asking a question about Arunachal Pradesh, the conflicted Indian state, but the model did not give an answer. 

However, this is not a problem with using DeepSeek on first-party mobile and web apps, not if you plan to deploy the model locally. The model’s open-source nature allows developers to remove restrictions and modify it according to their needs. 

Moreover, data collection concerns have also come to light. Luke de Pulford, a human rights activist, observed in DeepSeek’s privacy policy that it “collects your IP, keystroke patterns, device info…and stores it in China”. 

But again, the model can be downloaded and used locally, eliminating the concerns of data sharing via the internet. 

Another problem DeepSeek faced was a large-scale cyber attack. This led the company to temporarily stop onboarding new users. Moreover, it was also affected by outages on its website, but DeepSeek did resolve those issues soon after. 

Not Done Yet

That’s about all the major developments DeepSeek has made in the past few days, but the company isn’t done yet. It has announced Janus Pro, an AI image generation model that is claimed to offer better results than OpenAI’s DALL-E 3. 

It will only get interesting if DeepSeek enters the video generation competition – so Sora and Google Veo, better watch out! 

The post DeepSeek’s Dramatic Dominance: Everything You Need to Know appeared first on Analytics India Magazine.

As the Chinese AI lab DeepSeek gains global attention, so does the speculation about its founder. Misinformation has spread so widely that a fake account on X claiming to belong to CEO Liang Wenfeng prompted the company to issue a public clarification.

The launch of ChatGPT sparked a generative AI revolution in 2022, and now, DeepSeek’s release is generating a similar reaction. DeepSeek has surged to the top of the Google Play Store, just days after its chatbot app secured the No. 1 spot on the Apple App Store.

When asked about being called the ‘industry’s catfish’ in an interview originally published by 36Kr—a reference to a Chinese TV show where catfish symbolize market disruptors due to their cannibalistic nature—Wenfeng responded, “We didn’t set out to be a ‘catfish’ stirring up the industry. It just happened by accident.”

But who is he, really?

Born in the Chinese province of Guangdong in the 1980s, Wenfeng grew up as the son of a primary school teacher before carving out a career in finance and AI. According to reports, he launched investment firm Jacobi in 2013 and later founded High-Flyer, a Chinese hedge fund company, in 2015. The company now manages $8 billion in assets.

He was reportedly inspired by Jim Simons, founder of hedge fund Renaissance Technologies, for his mathematical approach to financial markets. He even wrote the introduction to the Chinese edition of The Man Who Solved the Market: How Jim Simons Launched the Quant Revolution, a book about Simons’ work.

In a way, he wants to achieve artificial general intelligence (AGI) to predict financial markets.

Han Xiao, CEO of JinaAI, described Wenfeng as a “low-key smart guy with no ego who constantly keeps learning and never wastes time on public exposure”.

Notably, the stock market crashed on Monday. Tech giants like NVIDIA, Microsoft, and Alphabet saw their stock prices plunge, while the Nasdaq 100 and Europe’s Stoxx 600 tech sub-index bled nearly $1 trillion in market cap. 

In the interview, Wenfeng argued that China’s AI progress wasn’t limited by money but by a lack of confidence and a structured ecosystem to drive research – something Silicon Valley has mastered. “More funding doesn’t necessarily yield more innovation,” he said. 

Wenfeng also contextualised this with the example of ChatGPT. “When ChatGPT first launched, the general sentiment in China – from investors to major tech firms – was that the gap was too big, so it was better to focus on applications. But innovation requires self-confidence.”

What, perhaps, makes the model stand out even more is its open-source nature. Unlike many AI founders, Wenfeng has taken an approach to make his technology free and open source. “Open source is more of a culture rather than a commercial behaviour, and contributing to it earns us respect,” he said. He has also stated that DeepSeek has no plans to go closed source. 

Meta’s chief AI scientist Yan LeCun said, “To people who see the performance of DeepSeek and think China is surpassing the US in AI, you are reading this wrong. The correct reading is: Open source models are surpassing proprietary ones.”

LeCun added that DeepSeek’s success stems from open research and tools like PyTorch and Llama, which were developed by Meta. He believes that by building on existing work and contributing back, DeepSeek has shown the power of open-source collaboration in advancing the AI ecosystem. 

Wenfeng said DeepSeek is focused on AGI, which requires developing new model structures for better performance with fewer resources. “If your goal is purely to build an application, then reusing the Llama structure and quickly pushing out a product is a reasonable choice. But our goal is AGI, which demands deeper innovation,” he explained.

Moreover, Hugging Face is replicating the entire DeepSeek R1 pipeline for the open-source community.

However, while DeepSeek’s model is open-source, its training data remains unclear. On Fox News, David Sacks, US President Donald Trump’s AI advisor, claimed there is “substantial evidence that DeepSeek distilled knowledge from OpenAI”. The US Navy has also reportedly banned the use of DeepSeek’s AI, citing “potential security and ethical concerns”. According to Reuters, the National Security Council, the US body that advises the president on foreign policy and national security, is reviewing the implications of DeepSeek’s apps, which topped the Apple App Store this past weekend.

The post Who is DeepSeek CEO Liang Wenfeng? appeared first on Analytics India Magazine.

Chinese AI company DeepSeek has not only disrupted markets and shaken technologists worldwide but also raised concerns over its privacy policy. 

Rajeev Chandrasekhar, former Indian IT minister, took to X to question if DeepSeek was on the path to becoming the next TikTok. His remark hinted at the growing concerns about DeepSeek AI’s potential impact on user data and its broader geopolitical implications.

Responses to his post ranged from mild concerns to those banishing it. Devilal Sharma, an alumnus of IIT Madras, said that since the model is open source, it can be used locally without an internet connection. “Deploy it on your own servers inside your own country, and the data won’t go anywhere,” he said. 

Another opined that this is nothing different from what ChatGPT or any other AI model is doing. “Any company which uses user data and key information, stores in their country may at any time in future make us vulnerable as [a] consumer,” he said. 

In its privacy policy, the company mentions that it collects network connection information, which includes “your device model, operating system, keystroke patterns or rhythms, IP address, and system language”.

“The personal information we collect from you may be stored on a server located outside of the country where you live. We store the information we collect in secure servers located in the People’s Republic of China,” read another section of the privacy policy. 

The mention of data being stored in China will certainly raise eyebrows in the Indian government. In 2020, India banned over 50 Chinese applications, including TikTok, which had approximately 200 million users in the country. 

‘World Doesn’t Need’ a TikTok of AI

Five years ago, India’s IT ministry banned these apps due to ‘raging concerns’ about data security and the privacy of the Indian population. The ministry said that it received multiple complaints from citizens, the Indian Cyber Crime Coordination Centre, and the Computer Emergency Response Team regarding the same. 

“On the basis of these and upon receiving recent credible inputs that such apps pose a threat to the sovereignty and integrity of India, the Government of India has decided to disallow the usage of certain apps, used in both mobile and non-mobile internet-enabled devices,” read a press release from the ministry in 2020. 

These Chinese apps were also said to be banned due to the escalating tensions between India and China. These apps are also barred from government devices in countries like Australia, Belgium, Canada, New Zealand, and a few others. 

Recently, TikTok’s service was also suspended in the United States, owing to a bill signed into law by former American president Joe Biden. Soon after, when Donald Trump took charge, he granted an extension, thereby halting the enforcement of the ban for 75 days. 

“Only safe and trusted AI should be offered to/be available on the global internet to consumers,” Chandrasekhar cautioned. 

“The world doesn’t need and must not enable a TikTok of AI,” he added. 

Will DeepSeek Be Banned in India?

Several users are afraid that DeepSeek might meet a fate similar to TikTok. “One of my worst fears right now is the Indian government banning DeepSeek for being a Chinese product,” said Shlok Khemani, a software developer on X.

Interestingly, the United States Navy has already banned DeepSeek and issued a warning to its members to avoid using DeepSeek for any purpose, whether work or personal. 

DeepSeek is also under scrutiny for avoiding controversial questions about China and for not answering questions about Arunachal Pradesh, an Indian state that is in conflict with China. 

India is likely to take a nuanced approach to dealing with any concerns posed by DeepSeek. A few days ago, the IT ministry drafted the ‘Digital Personal Data Protection Rules 2025’ in order to implement the ‘Digital Personal Data Protection Act (DPDP)’ from 2023. 

“It aims to strengthen the legal framework for the protection of digital personal data by providing necessary details and an actionable framework”, read a press release from the ministry. 

Last year, Chandrasekhar also criticised Google’s AI after it carried negative remarks about India’s Prime Minister. Soon after, the government issued an advisory requiring AI firms to obtain prior permission to make their products available online. However, another advisory soon followed, suppressing the previous one. 

Meanwhile, Chandrasekhar, while a tad sceptical, was also full of praise for the AI model. “This [is] a Chinese company demolishing the assumption that progress in AI and innovation in AI somehow depends upon billions of dollars of compute infrastructure.”

He indicated that this was against the narrative of AI development in the United States, and DeepSeek has shown the big western companies how powerful models can be built at a fraction of the cost. 

DeepSeek’s open-source nature allows it to be hosted on any server. For example, Perplexity offers DeepSeek-R1 on its platform and hosts the model on servers in the United States.

Meanwhile, there may be positive news on the Indo-China bilateral relationship, with both countries agreeing in principle to resume direct flights five years after the COVID-19 pandemic. The foreign ministry also said that the two countries will resume the Kailash Mansarovar Yatra in the summer of 2025. 

“The two sides agreed to take appropriate measures to further promote and facilitate people-to-people exchanges, including media and think-tank interactions,” read the announcement. 

The post ‘Is DeepSeek the TikTok of AI?’ Asks Former Indian IT Minister  appeared first on Analytics India Magazine.

As we celebrate Data Privacy Day, it is important to acknowledge the increasing complexity of cybersecurity and the higher stakes involved in protecting digital assets. 

During a recent session at the World Economic Forum (WEF), experts discussed the rise in cyberattacks and the urgent need for effective defences. “Whenever there is conflict in the political space, cyber conflict follows,” said Matthew Prince, CEO of Cloudflare. 

Highlighting the scale of the challenge, Prince noted that his company blocks over 220 billion cyberattacks daily. From Russia’s invasion of Ukraine to the conflict in Gaza, cyber warfare has targeted critical infrastructure and disrupted communication systems, underscoring the growing link between physical and digital conflicts.

The WEF Global Cybersecurity Forum 2025 report also highlights a concerning gap between organisations’ awareness of cybersecurity risks associated with AI and their preparedness to mitigate these risks. 

While two-thirds of companies recognise that AI will significantly impact cybersecurity this year, only 37% have the necessary tools to assess and address AI-related security threats effectively.

The challenge is particularly acute for smaller organisations, with 69% lacking adequate safeguards for the secure deployment of AI technologies. This lack of readiness could leave them vulnerable as AI adoption accelerates without sufficient protections in place.

A Collective Call to Safeguard Personal Data

From healthcare to commerce, education, and governance, safeguarding data has become essential to building trust and fostering innovation in every sector.

“At Dell Technologies, we believe data privacy is foundational to a resilient digital ecosystem. Organisations, governments, and individuals share a collective responsibility to protect sensitive information and uphold data rights,” said Ramesh Jampula, vice president, IT, India and APJC regional CIO, Dell Technologies.

Businesses Lead with Privacy-by-Design

Data privacy is no longer a box to check for compliance—it is a cornerstone of customer trust and financial stability.

“Our technology and operations platforms underpin the daily trading of, on average, more than $10 trillion in equities, fixed income, and other securities globally,” said Prasad Vemuri, chief information officer, Broadridge Financial India. 

“We recognize that robust data privacy transcends compliance—it is the bedrock of financial stability and client trust,” he said.

Meanwhile, SBM Offshore said that its advanced fleet of Floating Production Storage and Offloading (FPSOs) generate enormous volumes of operational data through extensive sensor networks that monitor equipment performance, safety, and efficiency. 

“We leverage this data using digital agents—advanced systems designed to detect anomalies and predict potential failures. These technologies enable us to take proactive measures and maintain high safety standards,” said Vikrant Sharma, IT department manager, SBM Offshore India.

Beyond laws and technology, Data Privacy Day is a reminder of the importance of empowering individuals to take control of their digital identities. Consent and control are central to ethical data practices, ensuring people can decide how their personal information is shared and used.

“This International Data Privacy Day, we should highlight consent’s vital role in digital ethical practices. When individuals share information willingly, they must maintain authority over its modification, deletion, and utilisation,” said Prashant Singh, COO, LeadSquared.

Commenting on the same thought, Jay Swamidass, vice president and global head of sales, Rakuten SixthSense said, “Privacy isn’t an upshot of chance. It comes from putting in the effort to create strong systems, holding ourselves accountable, and fostering a culture where privacy matters to everyone, at every level.”

He further mentioned that Rakuten SixthSense emphasises that privacy is fundamentally about trust—the trust individuals place in the organisation when sharing their data. Recognising that every piece of data represents a person, Rakuten SixthSense takes this responsibility with the utmost seriousness.

India’s Take on Data Privacy

India has made significant strides in data protection with the introduction of the Digital Personal Data Protection (DPDP) Act. The legislation places a strong emphasis on individual rights, such as consent, data correction, and erasure, while imposing strict penalties for non-compliance to prevent data misuse.

“The DPDP Act reflects the global push toward stricter data governance and privacy norms in an increasingly digital world,” said Rohith Reji, co-founder & CEO, Neokred. “It sets clear principles around lawful data processing, consent, and transparency.”

Commenting on the Draft Digital Personal Data Protection Rules 2025, a Reddit user wrote, “The million-dollar question is: What will happen to SPDI rules once the Digital Personal Data Protection (DPDP) Act comes into force?”

The user also said that unlike GDPR and other privacy laws, the Indian legislation does not further classify personal data within the Act itself, leaving open the question of how this aspect will evolve. “It will be interesting to see which direction it goes in,” as the lack of categorisation sets it apart from its global counterparts.

Another unique feature of the DPDP Act is its restriction to electronic data only. “It will be interesting to see how companies implement this limitation,” given its potential operational and compliance implications.

That said, this version of the Act appears far more refined than earlier drafts. “I do like this version over the previous draft and the ridiculous suggestions provided by the JPC committee,” such as the inclusion of non-personal data. These improvements reflect a more focused and practical approach to data protection in India.

Collaboration and Education

Strengthening privacy frameworks requires collaboration between industries, governments, and individuals. Companies like Bosch Global Software Technologies are championing continuous education and accountability to build a secure digital future.

“At Bosch Global Software Technologies, we recognize that data protection is not just about regulatory compliance—it’s about fostering trust and resilience,” said Vindhya Kudva, head of data protection & information security. “Through continuous education and partnerships, businesses can navigate data security complexities responsibly.”

Kudva added, quoting Zig Ziglar, “People often say motivation doesn’t last. Neither does bathing—that’s why we recommend it daily. The same applies to data privacy awareness—continuous education and vigilance are essential.”

Building a Privacy-Focused Future

The collective focus on reskilling, recruitment, and workforce development will be critical to building a privacy-focused digital future.

“As advancements in AI, edge computing, and IoT create new opportunities and challenges, the importance of data privacy will only grow,” said Jampula.

However, according to the WEF Global Cybersecurity Forum 2025 report, the sector faces a significant shortfall in cybersecurity talent, with up to 4.8 million professionals needed to bridge the gap. 

Only 14% of organisations report having the skilled workforce they require to tackle current cybersecurity challenges. In the public sector, the issue is even more pronounced, with nearly half (49%) of respondents stating they lack the workforce necessary to meet their cybersecurity objectives. The cyber skills gap increased by 8% in 2024.

“It’s critical we help close the growing cyber skills gap with a focus on training, reskilling, recruiting, and retaining cybersecurity talent,” said Chuck Robbins, chair and chief executive officer of Cisco.

Resonating the same, Publicis Sapient also recognises that safeguarding privacy requires more than just technology—it demands a workforce equipped with advanced skills and a mindset of continuous learning. 

“By investing in reskilling initiatives, particularly for AI engineers and cybersecurity experts, we aim to empower our people to design solutions that embed privacy by design and anticipate emerging risks. This culture of continuous learning ensures that our people are not just problem-solvers but forward-thinkers who can address tomorrow’s challenges today,” said Amit Patil, senior director of technology at Publicis Sapient.

The post Data Privacy Day: How GCCs and Indian IT are Leading the Charge appeared first on Analytics India Magazine.

Karan Kukreti began his career as a software developer but soon found himself yearning for something more fulfilling. The monotony of coding didn’t ignite his passion, but he discovered his true calling in analytics. He realised he was more drawn to the challenge of uncovering patterns in data and solving real-world business problems. However, pursuing his dream meant upgrading his skills and stepping out of his comfort zone. 

This was when he turned to Newton School, an edtech platform focused on providing real-world skills. Eventually, Karan successfully transitioned to a career in analytics and landed the role of a data analyst at Meesho. 

Newton School is a Bengaluru-based edtech startup founded in 2019 by Nishant Chandra and Siddharth Maheshwari. The four-year-old startup is a neo-university focused on empowering the next generation of tech leaders and entrepreneurs. 

In an exclusive interview with AIM, Chandra said, “The edtech startup’s journey begins with the background of the co-founders. Both Siddharth and I come from middle-class families and small towns. My parents migrated from Bihar to Delhi for better opportunities. My father was a government servant, and Siddharth’s family moved from a small town in Uttar Pradesh for similar reasons.”

This upbringing shaped their perspective. They attended IIT Roorkee, where they received excellent education and guidance – opportunities not readily accessible in many parts of India. 

“Even during college, Siddharth and I were involved in teaching students in nearby villages. Our mission has always been about democratising access to outcomes. Getting a job, especially for students from non-tier-1 colleges, is often luck-driven and uncertain. We wanted to change that,” he mentioned. 

Initially, they launched online programs, including pay-after-placement models, which saw massive success, with participation from students from over 500 colleges. They continued to focus on outcomes, ensuring that learning translates directly into employment. 

This Education Platform is for You

Between 2018 and 2019, the popular ed-tech players focused on K-12 education or competitive exams, such as Byju’s or Unacademy. These markets were quite distinct. Meanwhile, in the skilling segment, platforms like Coursera or Udemy provided learning materials but weren’t outcome-oriented. Their promise was more about self-paced learning rather than guaranteeing job placements.

“From the outset, we were focused on outcomes. We designed cohort-based live learning programs to build strong communities where students could learn and grow together. This approach addressed the high dropout rates in platforms like Coursera, where only 4-5% of students typically complete courses,” Chandra explained. 

He mentioned that factors like Jio’s affordable internet access and the normalisation of live online learning due to COVID-19 played a crucial role. They used live and interactive classes, which made learning much more engaging and effective. 

AI and Education 

According to a report by the National Skill Development Corporation (NSDC), 10 million students in rural India are using AI-powered learning tools. Meanwhile, 61% of educators are adopting AI tools for teaching, and student management, according to another report by TeamLease EdTech in January. 

India’s burgeoning edtech market has witnessed the AI transformation, and according to a LinkedIn report, the demand for AI skills in India is expected to grow by 50% in the next five years. The integration of technology into education presents both empowerment and cautionary tales. Personalised learning, real-time feedback, task automation, interactive experiences, and global access to resources are perfect examples of a transformative landscape.

“We started investing in AI in 2022, even before tools like ChatGPT became mainstream. Initially, using GPT-3 APIs, we developed tools for doubt-solving, mock interviews, and resume evaluation. Back then, accuracy was around 30-40%, and improving it to 80-85% would take months,” Chandra said.

However, with advancements like GPT-3.5 and GPT-4, they started at 80% accuracy. These models have been game-changers. For example, in September, their AI-powered doubt-solving tool handled over 1 lakh queries – a feat that would be impossible for human instructors alone.

“We also have AI-powered mock interviewers, which have helped students overcome their fear of interviews. One student, for instance, used the AI interviewer to practice over 15 sessions, which built her confidence for real interviews. Other tools, like AI resume evaluators and skill assessments, ensure our students are well-prepared for the job market,” he explained.

Newton School’s Partnership and More

Newton School has forged impactful partnerships to drive education and technology initiatives. It conducted training sessions on generative AI for ISRO, sponsored the ICPC India Regionals, and contributed to the development and support of the ICPC website. 

Collaborating with the IIT Roorkee Foundation, the institution is working on a project to create a portal to aggregate scholarships, job opportunities, and resources for SC, ST, and OBC communities and improve visibility and accessibility for marginalised groups. 

In partnership with the Telangana government, Newton School supported the infrastructure setup for a coding school at Kasturba Gandhi Balika Vidyalaya in Gopalpet to allow class 9 students to learn Python coding through a three-month online course with interactive learning tools. 

Additionally, Newton School has collaborated with over 800 companies to facilitate hiring, mentorship, industry insights, and real-world projects, thereby making sure that students are equipped with practical skills and job-ready expertise.

So, ultimately the edtech startup aims to embrace AI to democratise access to high-quality and outcome-driven education.

The post Newton School Turns Small-Town Struggles into a Mission to Democratise Tech Education appeared first on Analytics India Magazine.

Artificial intelligence still has much to achieve. However, last year, it received significant recognition for advancing biological sciences, thanks to Google DeepMind and its AlphaFold protein structure prediction technology, which played a crucial role in winning the 2024 Nobel Prize in Chemistry. 

But that’s merely the initial step in a long, ambitious journey to leverage AI to advance life sciences. In a fireside chat at the World Economic Forum (WEF) in Davos 2025, Demis Hassabis, CEO of Google DeepMind, stated, “Ultimately, my dream would be to simulate a virtual cell.” 

In another interview, he stated that AlphaFold provided a static interpretation of a protein, emphasising that one can only ‘truly’ understand what is occurring if the dynamics and interactions between different components within a cell are considered. 

“A virtual cell project is about building an AI simulation of a full working cell,” he said, he noted, suggesting he might start with a simple organism like a yeast cell.

Virtual Cell May Arrive Five Years From Now

At the World Economic Forum, Hassabis was accompanied by Ardem Patapoutian, a distinguished molecular biologist who was awarded the Nobel Prize in Physics or Medicine in the year 2021. Patapoutian elaborated on the idea of a virtual cell, saying, “Seeing the cell as is and where things are, both with respect to the architecture and 3D space, is very, very interesting.” 

He also mentioned that this aids us in achieving a deeper understanding of human biology beyond elevated levels of protein expression. 

“Let’s say a certain protein is expressed very highly in a cell. But when you actually look at it and see it, all of it is localised at the tip of the neuron where something very specific is happening,” he added. 

Hassabis also stated that this could be crucial for understanding how a cell reacts to the injection of a specific nutrient or drug. At present, this process is carried out ‘painstakingly’ in a laboratory, but according to him, this project would enable researchers to perform it a million times faster and at a lower cost. 

However, he noted that a laboratory would still be required as a final step to validate the predictions and conduct clinical trials to assess the efficacy of the research. Hassabis mentioned that the virtual cell project may materialise within five years from now. 

That said, this is one of many instances where researchers have investigated the concept of a virtual cell. 

How to Build an AI Virtual Cell

In December of last year, Stanford University, along with several prominent global universities, released a research study examining how artificial intelligence can create a virtual cell.

The researchers outline three capabilities that an AI virtual cell should possess. It should be able to create a universal representation across species and cell types, help accurately predict cell functions, allow computer-based experiments to test hypotheses, and guide data collection to expand its capabilities. 

The proposed model will consist of two components: an AI model to interpret the biological systems and ‘Virtual Instruments’ to analyse and simulate the results. 

The researchers said that this has the potential to help cancer researchers study how mutations cause healthy cells to become malignant, developer biologists to explore how cell lineages change with disruptions, and microbiologists to examine the effects of viral infections on both the cells and their host organisms. 

However, while one question asks ‘how to build a virtual cell’, the authors consider the more important one ‘how to trust in their competence’. 

“To this end, a comprehensive and adaptable benchmarking framework will be needed,” said the authors. 

“It [AI virtual cell] must account for dynamic distributions that evolve due to environmental changes, infections, genetic variants, and other such factors causing distribution shifts,” they added. 

Moreover, the authors mentioned another challenge: addressing bias in the data used to train the model and ensuring its diversity. 

“As virtual cell efforts mature, the dialogue between the scientists who develop models, those who generate experimental data, and funding organisations must be further intensified,” said the authors. 

However, they also emphasise the importance of using open data resources and standardised data formats to ensure that the project contributes to the overall benefit of humanity. This would also require a healthy collaboration between investors, bioethics experts, regulators, academia, researchers, and every other stakeholder involved. 

All things considered, all of these ambitions project towards increasing the longevity of life. 

Long Live..Longevity Tech?

“We don’t have an unambiguous metric for progress. If you must pick one, we can make a strong case for life expectancy,” reads a line from the book ‘The Anthology of Balaji’, a collection of statements made by Balaji Srinivasan, an entrepreneur who has built multiple genomics companies. 

Srinivasan opines that if the purpose of technology is to reduce scarcity, then the ultimate purpose of technology is to eliminate mortality, which is the main source of scarcity. 

Several technologists have embraced longevity technology. The recently announced $500 billion Project Stargate aims to eliminate diseases and improve longevity. 

According to Larry Ellison, CTO of Oracle and one of the companies funding the project, the infrastructure to be built aims to enable AI to “create cancer vaccines, personalised medicine, and pandemic prevention.” 

“We will see diseases get cured at an unprecedented rate. We will be amazed at how quickly we’re curing cancer and heart disease…” said Sam Altman, CEO of OpenAI, the company that is at the forefront of The Project Stargate. 

The company also has ambitions to build longevity tech, and Altman ‘personally funded’ $180 million to Retro Biosciences, a longevity science company, in 2023. Recently, Retro Biosciences also announced a $1 billion fundraising round to back clinical trials for three drugs. 

Most tech leaders share the same opinion. Dario Amodei, CEO of Anthropic, said, “If you think about what we might expect humans to accomplish in an area like biology in 100 years, I think a doubling of the human lifespan is not at all crazy,” while speaking at the WEF, in Davos. 

That said, we’re undoubtedly far away from realising such ambitions. In an interaction with AIM, Paras Chopra, founder of Turing’s Dream, who is also a gold medal recipient for his bachelor’s in biotechnology, believes that people are blowing up the ambition of immortality. 

“I’m all up for living longer and fitter, but I think we are up against very hard physical and biological limits when we’re talking about living forever,” said Chopra. 

“As technologists, we have a tendency to get super excited and super optimistic about things,” he added. 

The post Inside Google DeepMind’s Bold Vision for ‘Virtual Cell’  appeared first on Analytics India Magazine.