Yunjie Pan is a research scientist and PhD student in Computer Science at the University of Michigan with seven years of experience building high-performance ML systems and compilers. Currently at Meta, he focuses on productionizing cutting-edge training techniques—recent internships included fp8 mixed-precision training and PyTorch compiler work generating Triton code and autotuners for optimal layouts. His background spans industry internships at Twitter and deep academic training from Zhejiang University (top of his class) to advanced research in scalable model training. Known for bridging research and engineering, he combines compiler-level codegen expertise with practical autotuning to speed real-world large language model workflows. An attention to numerical efficiency and layout-aware implementations is a recurring, not-often-explicit thread in his work.
7 years of coding experience
Doctor of Philosophy - PhD, Computer Science and Engineering, Doctor of Philosophy - PhD, Computer Science and Engineering at University of Michigan
Bachelor of Engineering - BE, Electrical and Electronics Engineering, 3.99/4.0, Ranking 1/120, Bachelor of Engineering - BE, Electrical and Electronics Engineering, 3.99/4.0, Ranking 1/120 at Zhejiang University
A Python-level JIT compiler designed to make unmodified PyTorch programs faster.
Contributions:1 PR, 203 pushes, 49 branches in 3 months
pytorchpythonjitfastercompiler
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.