ADVERTISEMENTs
-10”-x-5.jpg)
Indian-origin AI expert exits Meta after five months
During his brief stint, Agarwal said his team advanced post-training techniques for “thinking” models, pushing an eight-billion parameter dense model close to Deepseek-R1
-
1749497506.png)
Malvika Choudhary - Wed, 27, Aug 2025
- 16:05 EDT
- 23
.png)
Rishabh Agarwal / X (@agarwl_)
Indian-origin AI researcher Rishabh Agarwal resigned from Meta just five months after joining its newly created Superintelligence Lab with a reported million-dollar salary package.
In a post on X, Agarwal announced, “This is my last week at @AIatMeta. It was a tough decision not to continue with the new Superintelligence TBD lab, especially given the talent and compute density. But after 7.5 years across Google Brain, DeepMind, and Meta, I felt the pull to take on a different kind of risk.”
Also Read: CalTech Prof. Anima Anandkumar to headline NACD Directors Summit
This is my last week at @AIatMeta. It was a tough decision not to continue with the new Superintelligence TBD lab, especially given the talent and compute density. But after 7.5 years across Google Brain, DeepMind, and Meta, I felt the pull to take on a different kind of risk.…
— Rishabh Agarwal (@agarwl_) August 25, 2025
He added that while the pitch from Meta CEO Mark Zuckerberg and chief AI scientist Alexandr Wang was “incredibly compelling,” he ultimately chose to follow Zuckerberg’s advice: “In a world that’s changing so fast, the biggest risk you can take is not taking any risk.”
During his brief stint, Agarwal said his team advanced post-training techniques for “thinking” models, pushing an eight-billion parameter dense model close to Deepseek-R1 performance, leveraging synthetic data to warm-start reinforcement learning, and improving on-policy distillation methods.
His exit comes at a turbulent time for Meta’s Superintelligence Lab, which has already seen multiple departures within months of its formation. According to reports, at least three other researchers, including Avi Verma and Ethan Knight, have also quit recently, with some returning to OpenAI.
Agarwal joined Meta in April 2025 after being recruited from Google DeepMind, where he worked on reinforcement learning, self-improvement, and distillation techniques for large language models.
An alumnus of IIT Bombay, he later earned a doctorate in artificial intelligence from Mila–Quebec AI Institute. His early career included stints at Saavn, Tower Research Capital, and Latent Logic (later acquired by Waymo).
He went on to spend five years at Google Brain, where his research in deep reinforcement learning won the NeurIPS 2021 Best Paper Award, before moving to DeepMind in 2023. He was also recently appointed adjunct professor at McGill University.
Related
.JPG)
.JPG)
.png)
.JPG)
.JPG)
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
E Paper
ADVERTISEMENT
.jpg)
Video
Comments
Start the conversation
Become a member of New India Abroad to start commenting.
Sign Up Now
Already have an account? Login