Qihang Yu
Research Scientist, Frontier AI & Robotics at Amazon
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolQihang Yu is a research scientist focused on computer vision and deep learning with nine years of experience applying these methods to medical image analysis and large-scale vision systems. Currently at Amazon Frontier AI & Robotics after research roles at TikTok and Google AI, he brings a mix of academic rigor from his PhD work at Johns Hopkins and hands-on engineering from internships at NVIDIA and Adobe. He has contributed to high-profile open-source projects like DeepLab2, improving data pipelines, optimizer support, and integrating advanced pixel encoders and panoptic augmentations, signaling a strong systems-to-models fluency. Comfortable moving between prototype research and production-ready code, he has a track record of shipping visualization and robustness improvements that benefit both model development and deployment. Based in San Francisco with dual backgrounds in CS and economics from Peking University, he combines technical depth with a pragmatic, cross-disciplinary perspective.
9 years of coding experience
7 years of employment as a software developer
Johns Hopkins UniversityBachelor's degree Economics, Bachelor's degree Economics at Peking University
Github Skills (5)
computer-vision10
deep-learning10
tensorflow10
machine-learning9
data-augmentation8
Programming languages (2)
Github contributions (5)
google-research/deeplab2
Jun 2021 - Jul 2022
DeepLab2 is a TensorFlow library for deep labeling, aiming to provide a unified and state-of-the-art TensorFlow codebase for dense pixel labeling tasks.
Role in this project:
ML Engineer
ML EngineerContributions:28 commits, 11 comments in 1 year
Contributions summary:Qihang primarily contributed to the DeepLab2 repository, a TensorFlow library for dense pixel labeling. Their work focused on improving visualization, fixing data loading issues, and adding support for the AdamW optimizer. They also integrated and added K-MaX's pixel encoder, decoder, and transformer decoder, as well as panoptic copy-paste augmentation, demonstrating an active role in enhancing the model's capabilities.
artdeep-learningtensorflow-librarystate-of-the-artaiming
yucornetto/MGMatting
Dec 2020 - Apr 2021
This repository includes the official project of Mask Guided (MG) Matting, presented in our paper: Mask Guided Matting via Progressive Refinement Network
Contributions:83 commits, 42 pushes, 28 comments in 4 months
pytorchmattingmaskrefinementdeep-learning
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.