Manan Tomar
Research Scientist at Microsoft
New York, New York, United States
Summary
👤
Senior🎓
Top SchoolManan Tomar is a research scientist at Microsoft Research NYC with a decade of experience building and advancing reinforcement learning and pretraining methods for vision and sequential decision-making. He completed a PhD at the University of Alberta after research stints at UC Berkeley’s RAIL lab and an AI Residency at Meta FAIR, where he developed policy optimization algorithms including Mirror Descent Policy Optimization. At MSR he focuses on latent prediction, jumpy-in-time reasoning, and discrete latent tokenization for smarter pretraining, blending theoretical RL insights with practical model engineering. His trajectory spans academic rigor and industrial research internships across Montreal and New York, reflecting strong collaborations with leaders like Sergey Levine and Philip Bachman. Notably, he has combined interests in image/video RL with work on tokenization and latent modeling, signaling a rare cross-cutting focus on both control and representation learning.
10 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy, Doctor of Philosophy at University of AlbertaIndian Institute of Technology Madras
English, German, Hindi
Github Skills (18)
gun10
swig10
reinforcement-learning9
algorithms9
cmake8
machine-learning8
openai8
artificial-intelligence8
data-science7
stable-baselines3
puzzle3
deep-reinforcement-learning3
sac3
autoencoder2
representation-learning2
Programming languages (3)
Github contributions (5)
manantomar/MPC
Jan 2018 - Feb 2018
Contributions:15 commits, 12 pushes, 1 branch in 1 month
manantomar/manantomar.github.io
Sep 2017 - Aug 2025
Contributions:69 pushes, 1 branch in 8 years
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.