From Kanpur to Meta: Tracking the meteroic rise of Indian-origin researcher Trapit Bansal; joins Zuckerberg's Superintelligence 11

Trapit Bansal , an Indian-origin AI researcher and co-creator of OpenAI ’s o-series models, has joined Meta as part of its newly launched Superintelligence Labs . A graduate of IIT Kanpur and the Indian Institute of Science (IISc), Bansal’s journey from academia to the forefront of global AI research underscores the rapid rise of Indian talent in Silicon Valley’s competitive landscape.

His move marks another high-profile departure from OpenAI to Meta, as Mark Zuckerberg intensifies efforts to dominate the race to artificial general intelligence (AGI).

A stellar academic foundation

• Undergraduate: Bansal studied at the Indian Institute of Technology (IIT) Kanpur.
• Postgraduate: He pursued research under Prof. Andrew McCallum at UMass Amherst, focusing on natural language processing (NLP), meta-learning, and reinforcement learning.
• Research Excellence: Best Paper Award at ICLR 2018 for work on meta-learning.
• Work featured in Wired, MIT Tech Review, and Business Insider.
• Contributed to competitive self-play research with OpenAI, showing how AI agents can acquire complex physical skills without reward engineering.

Research career across the big four
Bansal has built a rare résumé that includes stints at four of the most influential AI institutions globally:

OpenAI (2022–2025):

• Member of Technical Staff.
• Helped launch the RL for reasoning initiative with co-founder Ilya Sutskever.
• Contributed to OpenAI’s first reasoning model, o1.

Microsoft Research (2020):

• Worked on self-supervised meta-learning for NLP.

G oogle Research (2018)

• Focused on knowledge graph reasoning and deep learning.
  

Facebook (Meta) (2016):

• Developed deep learning models for NLP in the Applied Machine Learning group.

Joining Meta’s superintelligence push
Bansal joins Meta as one of 11 elite hires for its new Superintelligence Labs, led by Alexandr Wang of Scale AI and former GitHub CEO Nat Friedman. While his role remains undisclosed, Bansal’s expertise in reasoning models positions him as a crucial asset in Meta’s AGI ambitions.

“Thrilled to be joining @Meta! Superintelligence is now in sight,” Bansal wrote on X.

Thrilled to be joining @Meta! Superintelligence is now in sight 🚀

— Trapit Bansal (@TrapitBansal) June 30, 2025

Other notable hires include former OpenAI and Google talent such as Ji Lin, Hongyu Ren, and Huiwen Chang.

The Bigger Picture: Meta’s Billion-Dollar AI Gamble
Mark Zuckerberg is personally driving Meta’s AI transformation:

• Massive Spending: Meta spent over $14 billion to acquire a 49% stake in Scale AI, bringing in Wang to lead MSL.
• Aggressive Hiring: Reports suggest Meta offered bonuses up to $100 million to lure OpenAI engineers.
• Talent Exodus: OpenAI executives are scrambling to retain researchers, with Chief Research Officer Mark Chen describing the loss of staff as akin to a “break-in”.

Industry response: Mixed reactions
While Meta’s stock continues to climb, some experts question the sustainability of its strategy.

“There are some extreme downsides to going pure mercenary,” tech blogger Zvi Moshowitz told AFP.

“I don't expect it to work, but I suppose Llama will suck less.”

Institutional investors are also wary. Baird strategist Ted Mortonson warned that Zuckerberg’s unchecked spending could create long-term financial risks.

A new era beckons
Despite doubts, Meta’s long-term vision for AI is clear: to develop systems that surpass human reasoning, potentially revolutionising how content, ads, and products are created and delivered.