Yuchen Jin
Co-Founder And CTO at Hyperbolic
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolYuchen Jin is a Co-Founder and CTO based in San Francisco with seven years of experience bridging systems research and production ML infrastructure, and a PhD in Computer Science from the University of Washington. He led engineering work at OctoML developing Relax, a next-generation graph language that was merged into Apache TVM Unity, and now focuses on productizing compiler and deployment technology at his startup. An active open-source contributor, he has made deep technical contributions to apache/tvm (Relax VM codegen, shape lowering, Python array API/DLPack support) and to mlc-llm (LLM deployment ergonomics and Rust ChatModule support). His background includes research internships at ByteDance and Microsoft, producing practical systems like an automatic learning rate scheduler and a deployed latency-localization tool. He’s known for pragmatic, detail-oriented engineering that improves interoperability and usability across model runtimes and serving stacks—down to API ergonomics and real-world edge cases.
7 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of WashingtonBachelor's degree, Computer Science, Bachelor's degree, Computer Science at Huazhong University of Science and Technology
Github Skills (21)
c-language10
develop10
language-model10
compilation10
python10
machine-learning10
language-models10
llm10
compiler-compiler10
vms10
compile10
cuda10
compiler10
rust10
c-programming-language10
Programming languages (5)
Github contributions (5)
apache/tvm
May 2021 - Jul 2021
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:45 reviews, 10 commits, 26 PRs in 3 months
Contributions summary:Yuchen primarily contributed to the development of the Relax VM, a core component of the TVM project, focusing on its codegen and shape lowering aspects. Their work involved implementing support for the new Python array API with DLPack and addressing documentation issues, indicating a focus on improving interoperability and usability. Furthermore, the user refactored and updated the codebase, including renaming "gpu" to "cuda," integrating new VM functionalities, and optimizing code generation, demonstrating proficiency in compiler development within the context of machine learning.
metalvulkancompilertensoropencl
mlc-ai/mlc-llm
Sep 2023 - Jan 2024
Universal LLM Deployment Engine with ML Compilation
Role in this project:ML Engineer
ML EngineerContributions:8 PRs, 8 comments in 3 months
Contributions summary:Yuchen primarily contributes to the project by enhancing the `ChatModule` functionality, which is essential for the deployment of LLMs. Their work involves debugging the emoji handling in the REST API, indicating a focus on practical model serving and usability. The user also introduced and expanded Rust support within the repository, including the addition of Rust-based ChatModule interface, further demonstrating a contribution in LLM deployment. Furthermore, they modified the ChatModule API to support multiple prompts and improved its overall ergonomics.
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.