Wei Wei
Software Engineer at Meta
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolWei Wei is a software engineer with eight years of experience specializing in AI/ML infrastructure, engineering optimization, benchmarking, and deep learning systems. Currently at Meta after roles at Apple and Alibaba, he brings deep practical expertise in GPU-accelerated ML stacks—working on PyTorch internals like the AOT Inductor and the FX2TRT compiler for NVIDIA TensorRT to optimize operators, dynamic shapes, and XL-weight support. His background spans GPU architecture simulation and cloud AI infrastructure, giving him a rare combination of low-level hardware-aware engineering and high-level ML systems development. Based in Menlo Park, he pairs a PhD-level electrical engineering foundation with hands-on open-source contributions to flagship projects in the PyTorch ecosystem, evidencing both research depth and production-grade delivery.
8 years of coding experience
9 years of employment as a software developer
Postdoc, Postdoc at University of MichiganBachelor of Engineering - BE, Electrical Ennineering, Bachelor of Engineering - BE, Electrical Ennineering at Northwestern Polytechnical UniversityPhD, Electrical Engineering, PhD, Electrical Engineering at Mississippi State University
Github Skills (12)
tensorrt10
cuda10
pytorch10
machine-learning10
c-language10
tensor10
deep-learning10
cprogramming-language10
python10
gpu9
autograd8
open-source7
Programming languages (4)
Github contributions (5)
pytorch/TensorRT
Mar 2022 - Jan 2023
PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:56 reviews, 135 commits, 86 PRs in 10 months
Contributions summary:Wei primarily contributes to the PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT. Their work involves optimizing and fixing bugs within the FX2TRT framework, specifically in areas like operator support (e.g., `ne`, `isinf`, and `std`), reshaping, and handling of dynamic shapes. The user also contributes to the support of new operations, like supporting for XL weight. The user's contributions include writing unit tests, fixing existing issues, and improving the framework's overall capabilities.
cudapytorchjetsonnvidiadeep-learning
pytorch/pytorch
Jan 2022 - Sep 2022
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:ML Engineer
ML EngineerContributions:104 reviews, 24 commits, 42 PRs in 8 months
Contributions summary:Wei primarily contributed to the PyTorch library's optimization and performance, focusing on the AOT (Ahead-of-Time) Inductor. Their work included fixing hardcoded output data types, adding functionality for fuse_parallel_linear, and improving bandwidth profiling. They also addressed issues related to pre-grad passes within the AOT Inductor and made adjustments to the predispatch tracing API.
pythongpu-accelerationdeep-learninggpunumpy
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.