Jeff Niu
Member Of Technical Staff at OpenAI
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolJeff Niu is a compiler-focused software engineer with a decade of experience building high-performance systems and ML tooling, currently contributing to Triton at OpenAI after stints at Modular, Google (MLIR/TF/XLA), and Mojo work. He bridges low-level graphics and kernel experience with modern compiler backends, having fixed tricky MLIR/tablegen bugs and driven codegen and layout/performance improvements in projects like Triton and LLVM. Jeff is an active open-source contributor who also brings practical engineering chops from game-engine and HPC optimizations, demonstrating a knack for spotting and eliminating subtle performance and correctness issues. Based in the Bay Area and trained in mechatronics at the University of Waterloo, he’s a self-described “TensorFlow refugee married to MLIR,” signaling deep specialization in ML compiler stacks and a pragmatic taste for tooling that scales.
10 years of coding experience
6 years of employment as a software developer
Bachelor of Applied Science - BASc, Mechatronics Engineering, Bachelor of Applied Science - BASc, Mechatronics Engineering at University of Waterloo
Github Skills (35)
data-visualizations10
unit-testing10
javascript10
c-language10
dashboard10
compiler-development10
llvm10
data-visualisation10
mlr10
java10
codegen10
javas10
triton10
code-generation10
performance-optimization10
Programming languages (12)
Github contributions (5)
llvm/llvm-project
Apr 2020 - Jan 2026
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:165 reviews, 77 PRs, 334 pushes in 5 years 9 months
Contributions summary:Jeff primarily contributed to the MLIR compiler infrastructure, focusing on optimizing code generation for operations and properties. They fixed bugs related to attribute setters and getter generation within the MLIR tablegen tool. The user also worked on performance improvements by emitting inline getters and setters and refactoring the codebase. These changes demonstrate a focus on compiler optimization and code generation techniques.
compilerllvmtoolchain
triton-lang/triton
Nov 2024 - May 2026
Development repository for the Triton language and compiler
Role in this project:Backend Developer
Backend DeveloperContributions:564 reviews, 415 PRs, 738 pushes in 1 year 5 months
Contributions summary:Jeff contributed to the Triton language and compiler by fixing bugs and improving the codebase. Their contributions include addressing timeout issues in the `gtest_discover_tests` function, removing a workaround for an upstream bug in `DenseElementsAttr`, and generating local MLIR reproducers. The user also implemented a naive codegen for `tl.gather` and fixed shape mismatches in the for loop pipeliner. Additionally, the user refactored code, improved testing and implemented several layout and memory performance improvements within the compiler.
compiler
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.