Siyuan Feng
Assistant Professor at Shanghai Innovation Institute
Xuhui District, Shanghai, China
Summary
🤩
Rockstar🎓
Top SchoolSiyuan Feng is an assistant professor at Shanghai Innovation Institute and a PhD in Computer Science from Shanghai Jiao Tong University with nine years of experience in distributed machine learning systems and machine learning compilers. As an Apache Software Foundation member and PMC member for Apache TVM, he co-leads Relax, TensorIR, and TVMScript workstreams and has made substantive backend and CUDA/FP16/INT8 contributions to the widely used TVM compiler. He also contributes to MLC-LLM, focusing on ML compilation and model optimizations for GPU deployments, bridging research and production-grade tooling. Known for collaborating across academia and open-source communities (including close ties with Tianqi Chen), he combines rigorous research with hands-on engineering to push compiler-performance boundaries.
9 years of coding experience
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Shanghai Jiao Tong University
Chinese, English
Github Skills (11)
cuda10
compiler10
tvm10
code-generation10
language-model10
python10
llm10
cprogramming-language9
c-language9
webgpu8
vulkan5
Programming languages (11)
Github contributions (5)
mlc-ai/mlc-llm
May 2023 - Mar 2025
Universal LLM Deployment Engine with ML Compilation
Role in this project:
ML Engineer
ML EngineerContributions:54 reviews, 79 PRs, 93 pushes in 1 year 10 months
Contributions summary:Siyuan's contributions primarily focused on modifications and improvements to the ML compilation aspects of the project. The commits show the user fixing issues related to GPU parameters, addressing typos, and rebasing code related to the RWKV model. The user also updated documentation concerning model compilation for the RWKV architecture. The most significant contribution appears to be related to the compilation and optimization of various models within the MLC-LLM framework.
language-modelllmmachine-learning-compilationtvm
apache/tvm
Mar 2019 - Jan 2023
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:Back-end Developer & Compiler Engineer
Back-end Developer & Compiler EngineerContributions:4 releases, 912 reviews, 68 commits in 3 years 10 months
Contributions summary:Siyuan's commits primarily focus on enhancing the performance and functionality of a deep learning compiler stack. They are responsible for adding new features, providing support for the tensor core, and addressing compilation warnings in source files. They implemented and tested performance improvements for the CUDA backend, including support for FP16 and INT8, as well as modifications to the handling of memory allocations. Their work involves developing a new "while" node, enhancing the TVMScript parser and printer, and optimizing code for WebGPU.
metalvulkancompilertensoropencl
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.