Akshat Agarwal
Research Scientist At Meta Superintelligence Labs (MSL)
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolAkshat Agarwal is a research scientist and machine learning engineer with 11 years of experience, currently advancing multilingual and multicultural capabilities in Meta’s Llama and Muse models at Meta Superintelligence Labs. He bridges deep research and product delivery, with a background in robust multi-agent RL, graph neural networks, and production perception systems from roles at Carnegie Mellon and Nuro. Akshat has led teams building vision and sequence models for tracking, OOD detection, and trajectory prediction, then shifted back to hands-on model development and fine-tuning (SFT, RL) for localized instruction following, factuality, and formatting. He co-founded an award-winning open hardware startup (evive) that was crowdfunded and widely covered, reflecting a knack for turning research and prototypes into real-world impact. Based in San Francisco, he combines academic rigor with applied ML engineering across both research and system-level product constraints.
10 years of coding experience
8 years of employment as a software developer
Leaders In Innovation Fellow Entrepreneurship/Entrepreneurial Studies, Leaders In Innovation Fellow Entrepreneurship/Entrepreneurial Studies at Royal Academy of EngineeringIndian Institute of Technology KanpurMasters of Science - Research Robotics Artificial Intelligence, Masters of Science - Research Robotics Artificial Intelligence at Carnegie Mellon University
Hindi, English
Stackoverflow
176reputation
27kreached
2answers
1question
Github Skills (57)
kubernetes10
python10
rainbow10
gym-application10
docker-image9
reinforcement-learning9
benchmark9
gymnasium9
openai-gym9
fortress9
arxiv9
proximal-policy-optimization9
q-learning9
actor-critic8
docker8
Programming languages (5)
Github contributions (5)
agakshat/visualdialog-pytorch
May 2018 - May 2019
Contributions:17 commits, 8 pushes, 4 comments in 1 year
agakshat/spacefortress
Mar 2018 - Jun 2022
OpenAI Gym compatible reinforcement learning environment for Space Fortress https://arxiv.org/abs/1809.02206
Contributions:56 commits, 1 PR, 3 pushes in 4 years 4 months
rainbowdeep-reinforcement-learningbenchmarkreinforcementgym
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.