Gaurav Mahajan — AI Researcher

I’m a Postdoctoral Researcher at Yale’s Institute for Foundations of Data Science, where I work with Daniel Spielman. I earned my Ph.D. in Computer Science from UC San Diego (2023), advised by Sanjoy Dasgupta and Shachar Lovett. I’m transitioning from academia to industry, and looking for opportunities in post-training for language models. My research interests include reinforcement learning, post-training for language models, and learning theory: especially understanding the convergence properties of policy gradient methods (e.g. PPO, NPG, TRPO), and designing computationally efficient algorithms for learning language models. My work has appeared in venues such as COLT, ICML, NeurIPS, FOCS, ALT, and AISTATS, and spans topics including computationally efficient algorithms for learning language models (COLT 2023); the theory of policy gradient methods (COLT 2020); computational–statistical gaps in reinforcement learning (COLT 2022); and generalization frameworks for RL (ICML 2021, Neurips 2020, COLT 2020, FOCS 2021). Previously, I was a research intern at Microsoft Research, held visiting positions at the Institute for Advanced Study and the Simons Institute for the Theory of Computing, and was part of the Microsoft team that developed Microsoft PowerApps.

Stackforce AI infers this person is a Machine Learning Researcher with a focus on Reinforcement Learning and Language Models.

Location: New Haven, Connecticut, United States

Experience: 12 yrs 5 mos