Ryan Spring
Software Engineer, PyTorch at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolRyan Spring is a software engineer with 12 years of experience specializing in PyTorch backend development and CUDA code generation for high-performance deep learning. At NVIDIA he develops NvFuser and the PyTorch JIT compiler backend, researching ML-driven schedule synthesis to produce optimized GPU kernels. His open-source contributions to the flagship pytorch/pytorch repo include implementing elementwise ops and loss functions, reflecting deep hands-on expertise in tensor kernels and numerical primitives. Ryan’s academic background (PhD from Rice) bridges randomized hashing and large-scale distributed training, informing practical work on sparsity and scalable NLP systems. Based in California, he combines research rigor with production-grade engineering to push the frontier of compiler-driven acceleration for modern neural networks.
12 years of coding experience
1 year of employment as a software developer
Bachelor's Degree, Computer Science, Bachelor's Degree, Computer Science at The University of Texas at AustinDoctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at Rice University
Github Skills (11)
pytorch10
machine-learning10
tensor10
deep-learning10
python10
numpy9
autograd9
data-structure9
algorithm9
data-structures9
algorithms9
Programming languages (5)
Github contributions (5)
pytorch/pytorch
Nov 2020 - Dec 2022
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
Back-end Developer
Back-end DeveloperContributions:42 reviews, 77 commits, 21 PRs in 2 years 1 month
Contributions summary:Ryan's contributions primarily involve implementing and improving elementwise unary and binary operations, and loss functions within the PyTorch framework. They added reference implementations for various mathematical functions, including `nan_to_num`, `sigmoid`, `trunc`, `copysign`, `rsub`, `xlogy`, and more. Furthermore, the user incorporated loss functions like `mse_loss`, `l1_loss`, and `nll_loss`, extending the functionality of the library. The changes include modifications to the testing infrastructure and the inclusion of new primitives related to scalar values.
pythongpu-accelerationdeep-learninggpunumpy
rdspring1/comp450-planning_under_uncertainty
Oct 2014 - Feb 2019
Contributions:43 commits, 1 push in 4 years 4 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.