BharatGen

It was a privilege to host Abhay Karandikar at the BharatGen office.

Foundational AI for India cannot be imported. So how do you build it from scratch? China didn't get there by accident. It took a decade of research investment, an ecosystem where engineers and researchers worked toward the same problem, and a culture that treated compute efficiency as a scientific discipline. What can India learn from that? Prof. Ganesh sat down with Anirudh Suri on the AI Futures podcast to dig into exactly these questions, including the real constraints and how BharatGen is solving for them #beyondjugaad: - Data - Talent - Capital and compute And the bigger question: is India's participative, bottom-up approach to building AI a strength or a liability compared to the top-down models we see elsewhere? Tune in for the full conversation: 🎥 https://lnkd.in/ed78Q9A2

Congratulations to Prof. Ganesh Ramakrishnan on receiving the Dr. APJ Abdul Kalam HPC & AI Award 2026 on May 15. The award recognizes outstanding contributions in HPC and AI, specifically in large model development requiring high-performance compute. For Prof. Ganesh, this is not a new direction. It is the culmination of over two decades of work on building AI that performs in resource-constrained environments, with less data, less compute, and greater efficiency. That conviction, that constraints are a crucible for innovation, is what he brought to BharatGen. And it is what has made the work possible. We are proud and grateful to have him as our Founding Director. This recognition is well deserved, and it belongs equally to the researchers, engineers, and consortium members who have built alongside him. Department of Science & Technology, Government of India | IndiaAI Mission | Ministry of Electronics and Information Technology | Indian Institute of Technology, Bombay | Hewlett Packard Enterprise | NVIDIA | Rohan | Bhawna | RANGANATH | #HPEIndia #APJAbdulKalamHPCAwards

Thank you to the Tech Council of Australia and Monash University for the platform to share what BharatGen is building. It was wonderful to see Prof. Ganesh represent BharatGen at the event. Sovereign AI is a choice every nation is beginning to reckon with. We hope what we have learned building for India's languages and contexts is useful to others working through the same questions. Looking forward to more collaboration between our communities.

From Global Models to National Systems: Building India's Multilingual Large Language Model. We're hosting an important conversation in Melbourne on Friday 8 May with Monash University and we'd love to see you there. Our international guest, Professor Ganesh Ramakrishnan, leads BharatGen - one of the most advanced sovereign AI initiatives in the world. Built to serve India's languages, culture and infrastructure at national scale, BharatGen offers a compelling case study in what it actually means to build AI that reflects a nation's values and capabilities. As countries make consequential decisions about how to develop and deploy AI, the approaches they take will have lasting implications for productivity, investment, and trust. This session moves beyond the headlines to examine why India has taken the path it has and what that means for Australia. Joining Professor Ramakrishnan will be The Hon. Jaala Pulford GAICD, Dr Tim Rayner, and Professor James Bailey from Monash University. Register here: https://lnkd.in/ghgHecWX Friday, 8 May | 9.30am – 12.30pm Monash Conference Centre, 30 Collins St, Melbourne

At the Global Science Innovation Forum (GSIF) last week, Prof. Ganesh got to share some of the thinking that shapes how we build at BharatGen. A few of those threads: 1/ Provenance has to be in the stack from the beginning. Not added later. In how we collect data, in how we handle India's 22 scheduled languages and 1000+ dialects, in how research and engineering stay close enough that findings shape the pipeline in realtime rather than sitting in a paper. Because a system that cannot be traced cannot be trusted. This is as true for AI as it was for databases. 2/ We strive to be an ecosystem builder first. Models are a byproduct of that mission. The more important output is partnerships, shared infrastructure, digital public goods that others can build on. Project Tapestry, the federated open LLM initiative anchored by Yann LeCun's group, with BharatGen, EPFL, MBZUAI and others, is one expression of that thinking. You do not solve for the global south by building in isolation. 3/ On talent, the best advice we have received is to look in our backyard. The hunger of researchers coming out of IITs and India's academic institutions is real. Combined with the growing return of diaspora talent and the partnerships we are building through the IIT Bombay Heritage Foundation, the pipeline is stronger than it looks from the outside. Thank you to Prof. Anurag Mairal, PhD (He/His) and Megha Agrawal, Ph.D. at the Stanford Mussallem Center for Biodesign for putting together a genuinely thoughtful gathering. To Matt Chessen from RAND for moderating with the kind of rigour these conversations need. And to T.R. Vishwanath, co-founder and CTO of Glean, for a grounded and honest co-panelist perspective on enterprise AI governance. PS: We are hiring. If any of this resonates with how you think about building, we would love to hear from you. Photo credits: GSIF livestream on DiyaTV, YouTube

India's languages are not obstacles to building good AI. They are the architecture. At the "Transforming Tribal India Through Technology and Science" event organised by DST's SEED Division and implemented by North East Centre for Technology Application and Reach, NECTAR, Prof. Ganesh Ramakrishnan spoke about how BharatGen builds sovereign AI rooted in the structural unity across Indian languages, including tribal ones. The core argument: Indian languages share deep similarities in word order, case markers, and phonetics. That unity, when taken seriously at the model design level, means tribal and indigenous languages can be incorporated naturally into the same framework. No separate pipelines. No compromise on quality. BharatGen's models, across 22 Indian languages, built from first principles, are designed exactly this way. Grateful to DST and NECTAR for the platform to showcase our work.

BharatGen has signed an MoU with the Indian Institute of Technology, Bombay and CDACINDIA to advance foundational AI research for India. Signed on March 17, 2026, marking the 39th Foundation Day of C-DAC, this collaboration brings together three institutions to build sovereign AI infrastructure for India's linguistic diversity. C-DAC, an autonomous scientific society under the Ministry of Electronics and Information Technology brings deep expertise in multilingual computing, high-performance infrastructure, and technology solutions for socio-economic impact. These are areas that align closely with BharatGen's mission. Under this MoU, the three institutions will collaborate on: - Research on distributed AI infrastructure for foundational model development - Multimodal LLMs spanning speech, vision, and text - AI solutions for scheduled Indic, low-resource, and tribal languages - Data creation, model evaluation, and multilingual benchmark datasets India's linguistic diversity is one of the most complex challenges in AI, and one of the most important to get right. This collaboration brings together three institutions with a shared belief that sovereign AI for India must be built on a foundation of research excellence and national capability. From left to right: 1. Shashank Sharma, Scientist, C-DAC 2. Shri. Magesh E., DG, C-DAC 3. Shri. Sudeep Shrivastava, Joint Secretary, MeitY 4. Lenali Singh, Scientist, C-DAC 5. Prof. Rajat Moona, Director, Indian Institute of Technology Gandhinagar 6. Mahesh Bhargava, Scientist, C-DAC 7. samrit maity, Scientist, C-DAC 8. Dr Somanath, Former Chairman, ISRO - Indian Space Research Organization 9. Rishi Bal, CEO, BharatGen 10. Shri. Sanjay Wandhekar, Centre Head, C-DAC Pune 11. Prof. Ravi Kiran S., PI (Vision), BharatGen, International Institute of Information Technology Hyderabad (IIITH) 12. Dr. Maneesh Singh, VP of ML, BharatGen 13. pavan dhote, Scientist, C-DAC 14. Dr. Sheena Rani, DG, MED, CO & CS, DRDO 15. Dr. D. Srinivasa Reddy, Director, CSIR-INDIAN INSTITUTE OF CHEMICAL TECHNOLOGY

Building AI for a country as diverse as India requires depth in research and strength in deployment. At BharatGen, we have developed foundational AI models across text, speech, and vision for 22+ Indian languages. Our focus is on taking this research into real-world deployment. We are hiring for: • Text Engineering Lead • Speech Engineering Lead • Vision Engineering Lead • Eval Engineering Lead • ML Applied Scientist If this kind of work excites you, email us at careers@bharatgen.com