Papers
2
Total Citations
9
H-Index
2
About
Hanfei Yu is a rising researcher at the forefront of distributed systems and artificial intelligence, with a sharp focus on making deep reinforcement learning (DRL) both faster and more accessible. His work tackles a critical bottleneck in modern AI: the immense computational cost and time required to train DRL models for applications ranging from gaming AI and robotics to system scheduling. Yu’s major contribution lies in pioneering the use of serverless computing—a cloud paradigm that offers on-demand, elastic resources—to dramatically accelerate distributed DRL training. His highly cited 2024 paper, “Cheaper and Faster: Distributed Deep Reinforcement Learning with Serverless Computing” (7 citations), introduces novel architectures that slash training costs and time compared to traditional server-based clusters. Building on this, his follow-up work, “Nitro: Boosting Distributed Reinforcement Learning with Serverless Computing” (2 citations), further optimizes the actor-learner architecture for serverless environments. By demonstrating that serverless platforms can efficiently handle the trial-and-error, data-intensive nature of DRL, Yu is opening the door to more scalable and affordable AI development. His research is particularly notable for its practical impact, offering a blueprint for deploying advanced AI without expensive, dedicated hardware.
Research Focus
Key Achievements
Top Papers
- 1
- 2