Alex Nikulkov
Research Scientist at Meta
Seattle, Washington, United States
Summary
👤
Senior🎓
Top SchoolAlex Nikulkov is a research scientist with eight years of experience applying reinforcement learning, economics, and optimization to real-world product problems across Meta, Uber, and Eversight. Based in Seattle, he leads a small Data Science team and has shipped RL and policy-gradient components (including PPO and Balanced Off-Policy Evaluation) into the widely used ReAgent platform at Facebook Research. His background spans applied ML, experimentation, and operations-informed modeling—bringing together theory from a Stanford PhD in OIT and a near-perfect applied math degree from MIPT. He has a track record of turning research-grade algorithms into production-impacting systems like flexible delivery-fee models at Uber and decisioning platforms at Meta. Notably, he actively bridges AI/ML with operations management and economics, making him adept at solving both algorithmic and business-facing optimization problems.
8 years of coding experience
4 years of employment as a software developer
Ph.D., Operations, Information and Technology, Ph.D., Operations, Information and Technology at Stanford Graduate School of BusinessBS, Applied Mathematics and Physics, GPA 4.96 out of 5.0, BS, Applied Mathematics and Physics, GPA 4.96 out of 5.0 at Moscow Institute of Physics and Technology
Stackoverflow
71reputation
17kreached
0answers
1question
Github Skills (14)
model-building10
algorithms10
pytorch10
machine-learning10
bandit10
python10
reinforcement-learning10
modeling10
model-driven10
model-driven-development10
computer-engineering9
ppp9
hdfs6
hadoop6
Programming languages (2)
Github contributions (5)
facebookresearch/ReAgent
Jul 2019 - Sep 2022
A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:63 commits, 80 PRs, 1 comment in 3 years 1 month
Contributions summary:Alex implemented Balanced Off-Policy Evaluation (BOP-E) within the `reagent` platform, contributing to reinforcement learning and contextual bandits. They added a value model baseline and action masking support to the REINFORCE trainer, further enhancing policy gradient methods. The user also implemented a PPO trainer, indicating work on another core RL algorithm, and made several code improvements.
reinforcement-learningcontextualbanditscontextual-banditsreinforcement
alexnikulkov/ReAgent
Aug 2020 - Feb 2023
A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)
Contributions:120 pushes, 80 branches in 2 years 6 months
reinforcement-learningcontextualbanditscontextual-banditsreinforcement
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.