Muyang Li
Co-Founder And CEO at Stealth Startup
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMuyang Li is a PhD-trained engineer and founder with nine years of experience building and researching ML systems, currently leading a stealth startup as Co-Founder and CEO. Trained at MIT and Carnegie Mellon with a background in robotics and electrical engineering, he blends deep academic research with hands-on engineering from internships and research roles at NVIDIA and OmniML. His open-source contributions include practical tooling for GAN compression (CVPR 2020) where he implemented data-processing, evaluation, and latency measurement scripts—evidence of his focus on production-ready ML pipelines. He has a track record transitioning research into applied products, previously serving as a data scientist in industry and a visiting researcher at MIT. Based in Cambridge, he brings both entrepreneurial drive and low-level modeling expertise to scale interactive generative systems.
9 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Electrical and Electronics Engineering, Doctor of Philosophy - PhD, Electrical and Electronics Engineering at Massachusetts Institute of TechnologyMaster of Science - MS, Robotics, Master of Science - MS, Robotics at Carnegie Mellon UniversityBachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at Shanghai Jiao Tong University
Chinese, English, Japanese
Github Skills (7)
computer-vision10
pytorch10
python10
image-manipulation10
data-processing10
machine-learning9
compression8
Programming languages (7)
Github contributions (5)
mit-han-lab/gan-compression
Mar 2020 - Nov 2022
[CVPR 2020] GAN Compression: Efficient Architectures for Interactive Conditional GANs
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:7 releases, 112 commits, 2 PRs in 2 years 8 months
Contributions summary:Muyang primarily contributed to the project by adding and modifying Python scripts related to data processing and model evaluation. They created scripts for obtaining training IDs, preparing datasets for segmentation, and implementing image conversion and labeling functionalities. Furthermore, the user made contributions to testing the model by adding latency measurement and also made changes to the MUNIT model.
compressiongansconditional-ganspix2pixcyclegan
nunchaku-ai/nunchaku
Nov 2024 - Mar 2026
[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models
Contributions:8 releases, 58 reviews, 212 PRs in 1 year 4 months
diffusion-modelsfluxgenailoramlsys
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.