Wei Yang
Senior Research Scientist at NVIDIA
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolWei Yang is a Senior Research Scientist at NVIDIA Robotics Research Lab with 12 years of experience building and deploying deep learning systems for perception and robotics. He earned a Ph.D. in Electronic Engineering from The Chinese University of Hong Kong and spent a visiting scholar year at CMU working on visual navigation with leading researchers like Abhinav Gupta. His work spans model development and engineering—contributing substantial PyTorch implementations for classification and 2D human pose estimation, including adding modern architectures like pre-activated ResNet, WRN, DenseNet and ResNeXt. Based in Seattle, he blends research rigor with production-minded coding, often improving dataset support, data pipelines, and reproducibility in open-source toolkits. Notably, his contributions show a pattern of both advancing model capabilities and pragmatic engineering fixes that speed evaluation and integration.
12 years of coding experience
4 years of employment as a software developer
Bachelor of Software Engineering, Computer Software Engineering, Bachelor of Software Engineering, Computer Software Engineering at Sun Yat-Sen UniversityThe Chinese University of Hong Kong (CUHK)
English, Chinese
Github Skills (15)
neural-network10
mask-rcnn10
computer-vision10
faster-rcnn10
machine-learning10
pytorch10
deep-learning10
python10
pose-estimation10
classification10
human-pose-estimation10
resnet9
cifar1009
data-augmentation9
imagenet7
Programming languages (13)
Github contributions (5)
bearpaw/pytorch-pose
May 2017 - Dec 2021
A PyTorch toolkit for 2D Human Pose Estimation.
Role in this project:
ML Engineer
ML EngineerContributions:77 commits, 8 PRs, 50 pushes in 4 years 8 months
Contributions summary:Wei primarily contributed to the development of a PyTorch-based toolkit for 2D Human Pose Estimation. Their work includes implementing affine transformations, organizing code structure, and fixing bugs within the model and data loading components. The user also focused on saving predictions for evaluation, supporting different label map distributions and added new model. These changes improved the efficiency and functionality of the pose estimation pipeline.
pytorchhuman-posedeep-learningpose-estimationmscoco-keypoint
bearpaw/pytorch-classification
May 2017 - Jan 2019
Classification with PyTorch.
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:45 commits, 3 PRs, 11 pushes in 1 year 8 months
Contributions summary:Wei primarily contributed to the development of classification models using PyTorch. Their work involved the implementation of various neural network architectures, including Inception v3, and integration of a new CIFAR-100 dataset. Further, they were responsible for fixing bugs, improving code readability with "prettier print", and enhancing the model's capabilities. They also added pre-activated resnet, wide resnet (wrn), densenet and resnext support.
pytorchimagenetmobilenetpythonresnet
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.