Mikayel Samvelyan
Senior Research Scientist at Google DeepMind
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolMikayel Samvelyan is a Senior Research Scientist at Google DeepMind with eight years of experience at the intersection of foundational AI research and production-grade ML engineering. He previously contributed to FAIR at Meta and has a strong academic background including a PhD program at UCL and an MSc from Oxford, blending rigorous theory with practical system-building. His work spans research roles and hands-on engineering—examples include backend contributions to the widely used SMAC StarCraft multi-agent benchmark, where he improved environment mechanics and added heuristic AI and PySC2 support. Comfortable moving between research, teaching, and tooling, he brings a track record of shipping complex simulation and environment code that accelerates multi-agent and reinforcement learning experiments.
8 years of coding experience
9 years of employment as a software developer
Master of Science (MSc) Computer Science, Master of Science (MSc) Computer Science at University of OxfordMaster’s Degree Informatics and Applied Mathematics, Master’s Degree Informatics and Applied Mathematics at Yerevan State UniversityExchange Student Computer Information Systems, Exchange Student Computer Information Systems at Delta State UniversityUniversity College London
Armenian, English, Russian, German
Github Skills (13)
environmental10
dev-environment10
environ10
python10
reinforcement-learning10
enviroment10
benchmarking4
benchmark4
machine-learning4
data-structures4
algorithm4
algorithms4
data-structure4
Programming languages (5)
Github contributions (5)
oxwhirl/smac
Jan 2019 - Jan 2023
SMAC: The StarCraft Multi-Agent Challenge
Role in this project:
Backend Developer
Backend DeveloperContributions:12 reviews, 74 commits, 41 PRs in 4 years
Contributions summary:Mikayel primarily worked on the `StarCraft2Env` class, a core component of the SMAC environment, suggesting involvement in the back-end logic. Their contributions included importing necessary modules, initializing the game environment, defining actions, and handling rewards. The user also addressed issues related to the pathing grid and added the functionality for a heuristic AI, indicating a focus on environment mechanics and potential AI integration. The user further made modifications to enable support for PySC2 3.0.0 and added a new scenario.
starcraftstarcraft-iimultiagent-reinforcement-learningagentreinforcement-learning
facebookresearch/minihack
Sep 2021 - Jan 2023
MiniHack the Planet: A Sandbox for Open-Ended Reinforcement Learning Research
Contributions:6 reviews, 160 commits, 79 PRs in 1 year 3 months
multiagent-reinforcement-learningplanetstable-baselinespybulletreinforcement-learning
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.