James P. Gleeson

About

James P. Gleeson is a researcher whose work sits at the intersection of high-performance computing and machine learning systems, with a particular focus on understanding and optimizing deep reinforcement learning (DRL) workloads. His most-cited paper, "RL-Scope: Cross-Stack Profiling for Deep Reinforcement Learning Workloads" (2021, 4 citations), makes a foundational contribution by revealing that DRL workloads exhibit fundamentally different computational and memory-access patterns compared to traditional deep learning tasks. Gleeson’s key insight—that these structural differences create unique system-level bottlenecks—has helped shift how the community approaches DRL system design. By developing cross-stack profiling tools, he has enabled researchers and engineers to pinpoint inefficiencies spanning hardware, runtime, and algorithm layers. This work is particularly impactful for applications in robotics and data center management, where DRL’s promise is often limited by unpredictable performance. While his citation count is still growing, Gleeson’s contributions are notable for their clarity in diagnosing a previously overlooked problem, laying essential groundwork for future optimizations in reinforcement learning infrastructure.

Research Focus

Key Achievements

Top Papers

1. 1
   RL-Scope: Cross-Stack Profiling for Deep Reinforcement Learning Workloads

   4 citations · 2021

Key Collaborators

Contact & Links