Tingyu Wang
Senior Software Engineer at NVIDIA
United States
Summary
🤩
Rockstar🎓
Top SchoolTingwu Wang is a Staff Research Scientist at NVIDIA with 11 years of experience bridging machine learning research and real-world systems for games and robotics. He earned a PhD in machine learning from the University of Toronto with ties to Vector Institute and NVIDIA, and was advised by Sanja Fidler and Jimmy Ba. At NVIDIA he’s progressed from research intern to staff scientist working on generalist embodied agents and physics-based character animation using reinforcement learning. He’s an active open-source contributor in graph ML, integrating GPU-accelerated cugraph-ops into major libraries like PyTorch Geometric and DGL to speed up GNN layers such as SAGEConv, GATConv and RGCNConv. Based in California, he combines deep academic rigor with hands-on engineering to push scalable, GPU-first solutions for embodied AI. An under-the-radar strength is his practical ability to port research models into production-grade, high-performance code paths.
11 years of coding experience
Bachelor of Science (B.S.), Mechanical Engineering, Bachelor of Science (B.S.), Mechanical Engineering at Beihang UniversityDoctor of Philosophy (PhD), Engineering, Doctor of Philosophy (PhD), Engineering at The George Washington UniversityUniversity of Illinois Urbana-Champaign
Github Skills (9)
cuda10
pytorch10
deep-learning10
graph-neural-network10
gpu-acceleration10
graph-convolutional-networks10
python10
machine-learning8
geometric-deep-learning8
Programming languages (9)
Github contributions (5)
dmlc/dgl
Nov 2022 - Dec 2022
Python package built to ease deep learning on graph, on top of existing DL frameworks.
Role in this project:
ML Engineer
ML EngineerContributions:62 reviews, 3 commits, 7 PRs in 18 days
Contributions summary:Tingyu's primary contributions involve developing and integrating a custom `CuGraphRelGraphConv` model into the `dgl` library. This includes implementing the model, writing tests, updating documentation, and addressing review feedback. Furthermore, the user added an example that demonstrates the usage of the model for an entity classification task. The user also updated the `CuGraphRelGraphConv` module to use new bindings from `pylibcugraphops`.
pytorchpythondeep-learningmachine-learninggraph-neural-networks
pyg-team/pytorch_geometric
Dec 2022 - Aug 2023
Graph Neural Network Library for PyTorch
Role in this project:ML Engineer
ML EngineerContributions:1 review, 4 PRs, 10 comments in 7 months
Contributions summary:Tingyu contributed to the `pytorch_geometric` library by implementing and integrating optimized graph neural network models using the `cugraph-ops` package. Their primary focus was on creating GPU-accelerated versions of existing GNN layers, specifically `SAGEConv` and `GATConv`, leveraging the performance benefits of `cugraph-ops` for faster execution. They also worked on integrating RGCNConv model using cugraph-ops, which involved adapting the codebase to support the new functionalities.
pytorchgraph-convolutional-networksgeometric-deep-learningdeep-learningneural-graph
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.