Evgeny Bolotin
Papers
1
Total Citations
3
H-Index
1
About
Evgeny Bolotin’s research lies at the intersection of computer architecture, distributed systems, and machine learning, with a particular focus on the hardware-software co-design challenges of modern AI workloads. His work critically examines the architectural implications of running distributed reinforcement learning on heterogeneous CPU-GPU systems, revealing how scaling these algorithms exposes fundamental bottlenecks in memory bandwidth, synchronization, and power efficiency. By analyzing the performance and energy trade-offs of state-of-the-art RL training frameworks, Bolotin has provided essential guidance for designing next-generation accelerators that can support the growing computational demands of deep reinforcement learning. His contributions help bridge the gap between algorithmic advances in AI and the physical constraints of silicon, offering a roadmap for more efficient, scalable hardware architectures. Though early in its citation impact, his work on distributed RL systems is increasingly recognized as foundational for researchers aiming to deploy RL in real-world applications such as robotics, autonomous navigation, and game-playing agents. Bolotin’s research is a vital resource for students and engineers seeking to understand how future computing systems must evolve to keep pace with AI’s relentless progress.
Research Focus
Key Achievements
Top Papers
- 1