Hassam U. Sheikh

Orlando, Florida

I am a Software Engineer on the Reinforcement Learning team at Anyscale, where I am the technical owner of RLlib — the industry-standard distributed RL library. My work focuses on stability, performance, and correctness across RLlib’s distributed training and inference stack, including reliability hardening, benchmark-driven performance engineering, and regression-prevention guardrails at scale.

Previously, I was a Research Scientist at Intel Labs (2020–2024), where I published at ICML, ICLR, IJCNN, and AAMAS on multi-agent reinforcement learning, ensemble diversity, and intrinsic reward learning.

I hold a Ph.D. in Computer Science from the University of Central Florida, where I was advised by Ladislau Bölöni. My dissertation addressed stability challenges in multi-agent RL systems through the lens of defensive escort teams.

My research interests span distributed RL systems, multi-agent learning, ensemble methods, and offline RL.

news

Aug 01, 2025	Joined Anyscale as Software Engineer – RL. Now the technical owner of RLlib, the industry-standard distributed RL library.
May 15, 2022	Paper accepted at ICML 2022: DNS — Determinantal Point Process Based Neural Network Sampler for Ensemble RL.
Apr 10, 2022	Paper accepted at ICLR 2022: Maximizing Ensemble Diversity in Deep Reinforcement Learning.
Mar 01, 2022	Paper accepted at IJCNN 2022: Learning Intrinsic Symbolic Rewards in Reinforcement Learning.
Dec 15, 2020	PhD conferred in Computer Science from University of Central Florida. Dissertation: Multi-agent Reinforcement Learning for Defensive Escort Teams.