Chundian Li
Senior Software Engineer at Meta
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolChundian Li is a Senior Software Engineer based in Menlo Park with 11 years building high-throughput, scalable systems for billions of users. Currently at Meta (formerly Facebook), he has led notification recommendation infrastructure and driven privacy-aware platform work, pairing tech leadership with hands-on core development. His PhD training in distributed computing and micro-architecture informs a pragmatic approach to performance and correctness across large distributed stacks. An active contributor to PyTorch’s inductor component, he improves compilation/export paths and symbolic reasoning—bringing ML systems performance expertise into production infra. Known as “a realist with a perfection dream,” he balances rigorous engineering with practical shipping. He combines research-grade depth with product-focused delivery across both industry and open-source ML ecosystems.
11 years of coding experience
3 years of employment as a software developer
Bachelor's degree, Computer Software Engineering, Bachelor's degree, Computer Software Engineering at Wuhan UniversityDoctor of Philosophy (PhD), Distributed Computing, Micro Architecture, Doctor of Philosophy (PhD), Distributed Computing, Micro Architecture at University of Chinese Academy of SciencesVisiting scholar, Micro Architecture, Visiting scholar, Micro Architecture at Illinois Institute of Technology
English, Chinese
Github Skills (10)
pytorch10
machine-learning10
deep-learning10
python10
induction10
testing10
autograd9
neural-network9
tensor9
gpu8
Programming languages (3)
Github contributions (5)
pytorch/pytorch
Nov 2023 - Jan 2024
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
ML Engineer
ML EngineerContributions:15 reviews, 14 PRs, 128 pushes in 1 month
Contributions summary:Chundian primarily focuses on enhancing the PyTorch framework's performance and functionality, particularly within the `inductor` component. Their work involves fixing bugs related to `torch.split` and unbacked symints, optimizing symbolic guards for size variables, and adding test cases to prove non-strict export supports external calls. They also contribute to the export API by introducing tests and fixing issues related to non-strict mode. These contributions indicate a focus on improving the capabilities of PyTorch's compilation and export functionalities.
pythongpu-accelerationdeep-learninggpunumpy
andrewlee302/MIT-6.824
Nov 2016 - Nov 2017
Contributions:14 commits, 11 pushes, 2 branches in 1 year
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.