Boyuan Feng
Software Engineer at Meta
United States
Summary
🤩
Rockstar🎓
Top SchoolBoyuan Feng is a software engineer on the PyTorch Compiler team at Meta AI, bringing nine years of experience building deep learning systems and compilers. He holds a PhD from UC Santa Barbara where his research bridged high-performance computing and security for deep learning, with publications at venues like OSDI, ASPLOS, ATC, PPoPP, SC, and AAAI. Prior roles span research scientist work at Manta Ray Labs and machine learning internships at Facebook and Alibaba, reflecting a mix of industrial research and production-oriented engineering. Comfortable translating cutting-edge research into scalable tooling, he combines rigorous statistical training (MS from UW–Madison) and a math background from Nanjing University to optimize ML runtimes. Notably, his profile signals active engagement with both foundational research and the open-source ML ecosystem through his personal site and GitHub.
9 years of coding experience
2 years of employment as a software developer
Master of Science - MS, Statistics, Master of Science - MS, Statistics at University of Wisconsin-MadisonBachelor of Science - BS, MATHEMATICS AND STATISTICS, Bachelor of Science - BS, MATHEMATICS AND STATISTICS at Nanjing UniversityDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at UC Santa Barbara
Github Skills (9)
cuda10
transformer-models10
pytorch10
machine-learning10
tensor10
deep-learning10
python10
induction10
text-generation9
Programming languages (11)
Github contributions (5)
pytorch/pytorch
Apr 2025 - May 2026
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
Back-end & ML Engineer
Back-end & ML EngineerContributions:123 reviews, 77 PRs, 243 pushes in 1 year 1 month
Contributions summary:Boyuan's commits primarily involve refactoring and optimizing the PyTorch codebase, focusing on improving performance and adding new features related to tensor manipulation and deep learning. They made significant contributions to the development of the Flex Attention mechanism, including supporting various block mask configurations and addressing issues related to dynamic shape and numerical stability. The user also contributed to graph partitioning and related memory management strategies within the Inductor framework, as well as, handling edge cases for donated buffers.
gpu-accelerationneural-networkpythonautogradgpu
Manta-Network/wasm-zkp-challenge
Mar 2022 - Jul 2022
Contributions:33 commits, 3 PRs, 31 pushes in 4 months
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.