Xiaofeng Sun

Photo of Xiaofeng Sun

Xiaofeng Sun (孙啸峰)

Ph.D. Candidate

College of Computer Science and Technology

Zhejiang University, Hangzhou

Email: sunxf [at] zju [dot] edu [dot] cn

[Google Scholar] [GitHub] [Zhihu]

About me

I am currently a Ph.D. candidate at the College of Computer Science and Technology, Zhejiang University, supervised by Dr. hongyang Chen . Previously, I completed my Bachelor's and Master's degrees in Computer Science and Technology at China University of Petroleum (Beijing), where I was admitted to the integrated Bachelor-Master program with a ranking of 2/119.

News

  • Good News is cooming soon~

Research Interests

My research interests primarily lie in Large Language Model Optimization, Deep Learning Systems, and Efficient Model Inference. I am dedicated to addressing the performance and memory bottlenecks in AI model training and inference through algorithm-system co-design. I possess strong communication and teamwork skills, can independently analyze and solve problems, and have a keen interest in deep and challenging work.

Publications

  • Minimizing memory footprints for inference on IoTs devices by efficient rematerialization
    Xiaofeng Sun, etc.
    Published in the Proceedings of ICIC 2023 (CCF-C).
  • MEHTA: A Memory-Efficient Optimizer for Transformers with Hessian Trace Acceleration
    Xiaofeng Sun, etc.
    Manuscript in preparation, to be submitted soon.

Internship & Project Experience

  • Project: National Key R&D Program of China (Oct. 2022 - Sep. 2025)
    Key Technologies for AI Large Model Framework Optimization Based on Memory Pooling

Selected Awards and Honours

  • Outstanding Graduate of Beijing (2021)
  • First-class (2021) and Second-class (2022) Graduate Scholarships
  • Advanced Individual in the 70th National Day Parade (2019)