Sarvam AI Hires Former xAI Researcher Devendra Chaplot to Build Frontier AI Models

Former xAI, Thinking Machines Lab and Mistral AI researcher Devendra Singh Chaplot has joined Bengaluru-based Sarvam AI as an advisor, as the startup steps up work on frontier AI models in India.

The startup announced the appointment during its Epoch AI conference in Bengaluru, where the company unveiled new AI models and research initiatives. 

“I am excited to help Sarvam as an advisor and help them build the best models in the world,” Chaplot said at the event.

His appointment comes weeks after Sarvam raised $234 million in the first close of a $300 million Series B funding round at a $1.5 billion valuation. The company said the funding would support the development of its next-generation frontier AI models focused on agentic AI, coding, and cybersecurity, as well as investments in compute infrastructure and enterprise deployments.

The startup also reaffirmed that its previously announced one-trillion-parameter model remains under development and is expected within the next six months.

Pratyush Kumar, Co-Founder of Sarvam AI, said that Chaplot is advising the company on building large language models. The company has committed to developing trillion-parameter models, comparable in scale to those being built by leading Chinese AI firms, as it trains larger, more capable systems.

Chaplot brings experience from several of the world’s leading AI labs. Earlier this year, he joined Elon Musk’s xAI as a pre-training lead to work on AI research. Before that, he was a founding member of Paris-based Mistral AI, where he contributed to the development of models such as Mistral 7B, Mixtral 8x7B and Mistral Large.

Before joining xAI, Chaplot was an AI research scientist at Facebook AI Research (FAIR). He later moved to Thinking Machines Lab, where he served as a Member of Technical Staff.

Chaplot earned a PhD in Machine Learning and a Master's degree in Language Technologies from Carnegie Mellon University. He completed his BTech in Computer Science and Engineering at IIT Bombay.

Chaplot pushed back against the belief that India lacks the resources needed to develop frontier AI models.

“Many think building at the frontier requires resources that are impossible to get in India. That’s not true,” he said.

He further argued that building frontier AI does not depend on having a large number of seasoned researchers.

“You need a handful of experienced people, but mostly you need motivated and talented people. That’s something we have plenty of in India,” he said, adding that AI remains a relatively young field, allowing researchers to build expertise quickly with the right mentorship.

SarvamAI is also opening a new office in San Francisco. Kumar said the move is intended to attract top AI talent to work on building models for India.

“We want to have a very strong conduit of talent back to India. It’s important to get the best people thinking and building for the Indian context,” he said.

He further added that the company is already looking to hire in the US and is working through the logistics of setting up a US subsidiary and office.

Founded by Vivek Raghavan and Kumar in 2023, Sarvam AI develops models, inference infrastructure and enterprise AI products across speech, vision, document intelligence and AI agents. It was also selected by the Indian government under the IndiaAI Mission to build a sovereign large language model.