₹ 800 crore chase: Meta hires IITian Trapit Bansal from OpenAI

Trapit Bansal, an IIT-Kanpur alumnus and former OpenAI researcher, has joined Meta’s newly formed Superintelligence Labs, bringing with him deep expertise in deep learning, reasoning, and natural language processing to support Mark Zuckerberg’s ambitious AGI (artificial general intelligence) push.

Bansal’s professional journey reflects a rare blend of academic depth and industry exposure at some of the world’s top tech firms. A graduate in mathematics and statistics from IIT Kanpur, he later completed a PhD in computer science at the University of Massachusetts Amherst, specialising in meta-learning and NLP.

His career began in 2012 as an analyst at Accenture Management Consulting in Gurugram. He then moved into research at the Indian Institute of Science (IISc) Bangalore, focusing on Bayesian modelling and inference. Over time, he accumulated valuable experience through internships at Facebook, Google, Microsoft, and OpenAI, working on projects spanning deep learning for NLP, knowledge graphs, and reinforcement learning.

In 2022, Bansal joined OpenAI full-time, collaborating directly with co-founder Ilya Sutskever on reinforcement learning for reasoning. His work was instrumental in creating OpenAI’s first AI reasoning model, o1, which laid critical groundwork for ChatGPT’s internal reasoning capabilities.

Now at Meta Superintelligence Labs (MSL), a division launched by Zuckerberg to develop AGI under the leadership of Alexandr Wang (ex-Scale AI CEO) and Nat Friedman (ex-GitHub CEO), Bansal is expected to play a key role in advancing Meta’s next-generation large language models. With MSL aggressively hiring top AI researchers from OpenAI, Anthropic, and DeepMind, reports from Bloomberg and The Wall Street Journal suggest Meta has been dangling pay packages reaching $300 million over four years, including stock options and massive compute access. Though the exact terms of Bansal’s deal are not public, speculation points to similarly lucrative incentives.

This hiring spree has stirred tensions across the industry. OpenAI CEO Sam Altman recently remarked that Meta was offering "$100 million signing bonuses," though he claimed their best people had turned them down. Still, Bansal’s move, alongside exits by researchers like Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai, Jack Rae, and Johan Schalkwyk, suggests Meta’s overtures are proving hard to resist.

With Meta yet to unveil a reasoning-capable AI model, Bansal’s arrival is seen as crucial to closing that gap. His expertise in systems that can learn, reason, and plan is expected to directly influence Meta’s bid to build AGI and compete with OpenAI’s o3 and DeepSeek’s R1.

Meanwhile, despite Meta spokesperson Andy Stone downplaying reports of inflated salaries, industry insiders believe these hires underscore the urgency with which Meta is chasing its AGI ambitions.