Xiao Wang
Senior Software Engineer at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolXiao Wang is a Senior Software Engineer in California with a Ph.D. in theoretical/computational physics and eight years of experience applying rigorous math and systems thinking to ML infrastructure. At NVIDIA they work on PyTorch, TensorRT-LLM and agentic AI, focusing on CUDA-accelerated linear algebra and production-ready deep learning primitives. Xiao has contributed notable enhancements to the flagship pytorch/pytorch repo—adding a driver= kwarg for torch.linalg.svd/svdvals and enabling relocatable device code linking for CUDA extensions—to boost performance and build reliability. Equally comfortable in C, C++, Python and CUDA, they bridge low-level optimization with higher-level model tooling and have practical exposure to C#, Java and web stacks. Their background in computational physics informs a methodical approach to numerical correctness and performance engineering that few software engineers possess.
8 years of coding experience
Bachelor’s Degree, Physics, Bachelor’s Degree, Physics at University of Science and Technology of ChinaDoctor of Philosophy (Ph.D.), Physics, Doctor of Philosophy (Ph.D.), Physics at Purdue University
English, Chinese
Github Skills (12)
cuda10
pytorch10
deeplearning-ai10
deep-learning10
gpu10
cusolver10
python9
machine-learning9
tensor9
lg8
neural-network8
autograd8
Programming languages (13)
Github contributions (5)
pytorch/pytorch
Feb 2020 - Dec 2022
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
ML Engineer
ML EngineerContributions:197 reviews, 287 commits, 160 PRs in 2 years 10 months
Contributions summary:Xiao primarily contributed to the `pytorch/pytorch` repository by enhancing linear algebra and CUDA functionalities. They added a `driver=` kwarg to `torch.linalg.svd` and `svdvals`, integrating the cusolver gesvdaStridedBatched driver for improved performance. Additionally, the user implemented support for relocatable device code linking in CUDA extensions, improving the build process. These changes demonstrate a focus on optimizing and expanding the capabilities of PyTorch's linear algebra and CUDA integration, core components of its deep learning functionality.
gpu-accelerationneural-networkpythonautogradgpu
xwang233/code-snippet
Feb 2020 - Dec 2021
Contributions:150 commits, 134 pushes, 1 branch in 1 year 9 months
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.