π·οΈ Research Keywords
- GPU algorithms
- parallel algorithms
- irregular workloads
- systems for deep learning
- ML systems
- ML compilers
Final-year PhD candidate in Computer Science at the University of Toronto (UofT), advised by Prof. Gennady Pekhimenko. I build GPU and ML systems for irregular workloads, distributed inference, and compiler/runtime efficiency, with prior experience at AWS and ByteDance.
Amazon Web Services (AWS)
Applied Scientist Intern
- Optimized video generation workloads on AWS P4d machines by overlapping communication with computation, achieving up to 1.58Γ speedup.
ByteDance Inc.
Research Intern
- Explored how to integrate pretrained language models into neural machine translation by designing and evaluating multiple BERT knowledge distillation strategies.
- Optimized the training data pipeline.
University of Toronto
Research Assistant
- Conducted research in machine learning systems and high-performance systems under Prof. Gennady Pekhimenko.
- Representative work includes Minuet, ScaleFusion, and StreamFusion.
Massachusetts Institute of Technology (MIT)
Research Intern
- Conducted research under Prof. Song Han on reinforcement-learning-driven circuit automation.
- Representative work includes the paper Learning to Design Circuits.
Shanghai Jiao Tong University (SJTU)
Research Assistant
- Conducted early-stage research with Prof. Weinan Zhang and Prof. Yong Yu on reinforcement learning and learning-augmented systems.
- Representative work includes the many-agent reinforcement learning platform MAgent.
University of Toronto
Ph.D. in Computer Science
University of Toronto
M.A.Sc. in Computer Engineering
Shanghai Jiao Tong University (SJTU) Β· ACM Class
B.S. in Computer Science
Showing first-author papers by default. Switch to view the full list.
- ICML 2025, 2026 Reviewer
- NeurIPS 2024, 2026 Reviewer
- ACM CAIS 2026 Programming Committee
- MLSys Artifact Evaluation 2026 Reviewer
- EuroSys Artifact Evaluation 2025 Reviewer
- ICLR 2025 Reviewer