Jordan Fix
Research Scientist at Meta
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolJordan Fix is a Research Scientist in New York with eight years of experience building ML compilers and runtimes for hardware accelerators, currently contributing to PyTorch at Meta. He blends deep systems research—PhD work on transactional memory, ISA integration, and verifiable trusted components—with hands-on compiler engineering, having optimized neural-network compiler paths and added quantized operator support to the high-profile PyTorch/Glow ecosystem. Jordan’s background spans academic rigor at Princeton to production-focused tooling at Meta and Facebook, and his contributions include fusing ops (e.g., ReLU/RescaleQuantized) and extending ONNX model loading to broaden framework interoperability. He’s equally comfortable prototyping in simulators like gem5 and shipping performance-sensitive changes in large open-source compilers, a combination that surfaces in both research papers and tangible runtime wins.
8 years of coding experience
7 years of employment as a software developer
Masters, Computer Science, Masters, Computer Science at Princeton UniversityBachelor of Science, Computer Science, Bachelor of Science, Computer Science at University of Virginia
Github Skills (8)
compiler-optimization10
quantization10
c-language10
onnx10
cprogramming-language10
python9
tensorflow9
deep-learning9
Programming languages (3)
Github contributions (5)
pytorch/glow
Jan 2018 - Aug 2022
Compiler for Neural Network hardware accelerators
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:251 reviews, 721 commits, 728 PRs in 4 years 8 months
Contributions summary:Jordan's contributions focused on compiler optimizations for neural network hardware accelerators. They implemented performance improvements by merging and fusing operations like ReLU and RescaleQuantized nodes. The user also added support for quantized operations, including LayerNorm and Swish, and enhanced the functionality of existing nodes, such as the Scatter/Gather family and BatchMatMul. Furthermore, they contributed to extending ONNX model loading to support new operators, enhancing the Glow compiler's capability to execute models from diverse frameworks.
hardware-acceleratorscompilerneural-networkacceleratorshardware
jfix71/pytorch
Mar 2020 - Feb 2024
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:128 pushes, 49 branches in 4 years
pythongpu-accelerationdeep-learninggpuacceleration
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.