Sumeet Motwani

Oxford, United Kingdom

Bio

Updated 03/23/26

Sumeet Ramesh Motwani is a Machine Learning PhD student at the University of Oxford, advised by Philip Torr and Christian Schroeder de Witt, with funding from Eric Schmidt and CAIF. He completed his undergraduate degree in computer science at UC Berkeley, where he was a member of Berkeley AI Research (BAIR) advised by Dan Hendrycks. His research focuses on RL post-training, multi-agent systems, and AI security, with particular interests in meta-RL, open-endedness, and long-horizon LLM agent capabilities. He is known for his work on "Secret Collusion Among Generative AI Agents" (NeurIPS 2024), which established the subfield of secret collusion in multi-agent AI systems, and for "STARC: A General Framework For Quantifying Differences Between Reward Functions" (ICLR 2024), published while he was an undergraduate. He has also contributed to research on autonomous web-browsing agents (Agent Q, REAL benchmark) and multi-agent LLM training (MALT, COLM 2025). He participated in the MATS (ML Alignment Theory Scholars) program and is affiliated with the Future of Life Institute as a community researcher. He has held research positions at Microsoft Research (AI Frontiers lab) and Google X.