I am a final year PhD Student in the system research group at the Department of Computer Science and Technology, University of Cambridge, advised by Dr Eiko Yoneki. My research focuses on RL-driven performance optimization, and recently works on LLM systems for efficient training and inference.

I have worked in several industrial labs on building efficient LLM systems.

• Cohere model efficiency team: I worked on performance optimization of LLM serving.

• Bytedance Seed pre-training team: I studied the impact of physical network topology to pre-training LLMs at the scale of O(10,000) GPUs, and contributed to an internal scheduling system for pre-training LLMs.

• Shanghai AI Lab LLM infrastructure team: I contributed to LMDeploy, , a high performance inference framework for LLMs.

I have also contributed to open-source projects such as Pytorch and Triton. For more information, see my Project page.

Latest publications

(* denotes equal contributions)

LLM Systems, Training and Serving

• Demystifying Cost-Efficiency in LLM Serving over Heterogeneous GPUs
  ICML 2025
  Y. Jiang*, F. Fu*, X. Yao*, G. He*, X. Miao, A. Klimovic, B. Cui, B. Yuan, E. Yoneki

• CuAsmRL: Optimizing GPU SASS Schedules via Deep Reinforcement Learning
  CGO 2025
  G. He, E. Yoneki

• SIP: Autotuning GPU Native Schedules via Stochastic Instruction Perturbation
  EuroSys 2024, EuroMLSys workshop
  G. He, E. Yoneki

• vPALs: Towards Verified Performance-aware Learning System For Resource Management
  AAAI 2024, deployable AI workshop
  G. He, G. Yeung, S. Ceesay, and A. Barker

RL-Driven Performance Optimization

• Optimizing Tensor Computation Graphs with Equality Saturation and Monte Carlo Tree Search
  PACT 2024
  J. Hartmann, G. He and E. Yoneki

• X-RLflow: Graph Reinforcement Learning for Neural Network Subgraph Transformation
  MLSys 2023
  G. He, S. Parker, and E. Yoneki

• MCTS-GEB: Monte Carlo Tree Search is a Good E-graph Builder
  EuroSys 2023, EuroMLSys workshop
  G. He, Z. Singh, and E. Yoneki

Biography

• 2021 - 2025 : Studied PhD degree at the University of Cambridge.

• 2020 - 2021 : Worked as a software engineer in cloud infrastructure.

• 2019 - 2020 : Receive my M.Sc. degree in Machine Learning from University College London.

• 2015 - 2019 : Receive my B.Eng. degree from University of Edinburgh and South China University of Technology.