Kevin Gleason
Software Engineer (ML Compilers) at Google
Chicago, Illinois, United States
Summary
🤩
Rockstar🎓
Top SchoolKevin Gleason is a software engineer specializing in ML compilers with 12 years of experience, currently working on compiler infrastructure at Google from his base in Chicago. He focuses on MLIR/StableHLO integration and serialization across major projects—contributing to JAX, TensorFlow, XLA, and the StableHLO dialect—to enable dynamic-shape handling and deployable JAX programs on accelerators. His background includes years of compiler front-end design and language implementation at MathWorks, where he built AST transforms, parser tooling, and new MATLAB syntax features. An avid programming-language enthusiast, he combines practical engineering (bytecode, pretty-printing, and error-handling improvements) with performance-minded backend work on GPUs/TPUs. Outside work he’s a committed soccer fan, reflecting the same team-oriented, competitive mindset he brings to complex compiler problems.
12 years of coding experience
3 years of employment as a software developer
Gilmour AcademyBachelor of Science - BS, Bachelor of Science - BS at Boston College
Stackoverflow
1,032reputation
49kreached
48answers
1question
Github Skills (24)
c-language10
back-end-development10
python10
stable10
mlr10
tensorflow10
compiler-optimization10
xla10
jax10
cprogramming-language10
compiler-development9
machine-learning9
gpu-programming8
error-handling8
code-coverage7
Programming languages (10)
Github contributions (5)
openxla/stablehlo
Aug 2022 - Jan 2023
Backward compatible ML compute opset inspired by HLO/MHLO
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 release, 1481 reviews, 61 commits in 5 months
Contributions summary:Kevin's contributions focused on enhancing the StableHLO project by implementing pretty printing functionalities for various types within the StableHLO dialect. They added custom print directives and re-wrote parsing/printing functions. Furthermore, the user implemented bytecode implementations for new attributes and types, including OutputOperandAlias. They also implemented code coverage tooling. The work involved modifying and adding to `StablehloOps.cpp`, `StablehloBytecode.cpp`, and other relevant files.
hlomachine-learningcompute
openxla/xla
Dec 2022 - Jan 2023
A machine learning compiler for GPUs, CPUs, and ML accelerators
Role in this project:ML Engineer
ML EngineerContributions:33 reviews, 10 commits, 2 PRs in 1 month
Contributions summary:Kevin primarily contributed to the machine learning compiler's infrastructure, focusing on the MLIR dialect. Their commits addressed the pretty-printing of custom call operations, removed default types for StableHLO to HLO mappings to improve error handling, and replaced instances of `llvm::None` with `std::nullopt`. Additionally, the user added tests for error handling in the `printDenseI64Array` function and made code style improvements. These changes suggest a focus on enhancing the compiler's features and maintaining the codebase.
compilercommunity-drivenmachine-learningmodular
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.