Xiyou Zhou
Software Engineer at The Apache Software Foundation
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolXiyou Zhou is a software engineer with 13 years of experience specializing in ML compilers, performance optimization, and foundation models, currently working on Foundation Model efforts at Apple. A long-time Apache TVM committer and core contributor to MetaSchedule, he has driven automatic tuning and target configuration work that improves inference across NVIDIA/AMD GPUs and Intel/ARM CPUs. At OctoAI he led TVM-related LLM serving infrastructure and tools for model deployment, and his open-source contributions to projects like apache/tvm and mlc-ai/mlc-llm include CUDA target support, host-target handling, benchmarking tooling, and group quantization features. He blends research-grade rigor from internships at NVIDIA and Google with pragmatic engineering for production ML stacks. Based in Seattle, he pairs deep systems expertise with a knack for extracting portable workloads—evident in tooling that dumps PrimFuncs for cross-platform porting.
13 years of coding experience
4 years of employment as a software developer
Bachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at Fudan UniversityMaster of Science - MS Computer Science, Master of Science - MS Computer Science at UC Santa Barbara
Github Skills (14)
cuda10
tvm10
c-language10
language-model10
cprogramming-language10
python10
compilation10
compile10
llm10
compiler-compiler9
compiler9
gpu7
deep-learning7
machine-learning6
Programming languages (7)
Github contributions (5)
apache/tvm
Feb 2021 - Nov 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
Back-end Developer
Back-end DeveloperContributions:338 reviews, 75 commits, 136 PRs in 1 year 10 months
Contributions summary:Xiyou primarily contributed to the development of CUDA target configurations by adding CUDA tags, defining target configurations for various NVIDIA GPUs. The user also worked on features related to the target specification, including adding support for the host field within the target object. Furthermore, the commits reflect a focus on refining and improving the codebase, including applying formatting, resolving issues with target host and ensuring data type consistency.
metalvulkancompilertensoropencl
mlc-ai/mlc-llm
May 2023 - Apr 2024
Universal LLM Deployment Engine with ML Compilation
Role in this project:ML Engineer
ML EngineerContributions:16 reviews, 8 PRs, 22 comments in 11 months
Contributions summary:Xiyou made significant contributions to the ML compiler project, specifically focused on enhancing the benchmarking capabilities of the model deployment engine. They added scripts for dumping PrimFuncs in debug mode, facilitating the extraction and porting of workloads. Furthermore, the user addressed bugs related to sampling functions and enabled group quantization features, improving the model's performance and efficiency. They also enabled weight conversion functionality via the command line interface.
language-modelllmmachine-learning-compilationtvm
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.