Bart Chrzaszcz
Senior Software Engineer at Google DeepMind
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolBart Chrzaszcz is a Senior Software Engineer at DeepMind with nine years of experience building compilers and large-scale ML training infrastructure. He combines low-level systems expertise (LLVM/MLIR, JAX/OpenXLA) with practical build and release engineering, having fixed complex Bazel/MLIR build issues and implemented sharding features in JAX used for SPMD partitioning. His background spans model-scaling work on the Shardy team, early contributions to JIT and autodiff projects, and internships at Google Brain and Apple, reflecting a consistent focus on compiler-backed ML tooling. Based in London, he brings both research-adjacent engineering rigor and hands-on production experience that improves performance, correctness, and test coverage in widely used open-source projects. An underrated strength is his ability to bridge compiler internals and large-scale training pipelines—making low-level fixes that unlock scalable ML runtimes.
9 years of coding experience
5 years of employment as a software developer
Bachelor’s Degree Co-op Computer Science, Bachelor’s Degree Co-op Computer Science at University of Waterloo
English, Polish
Stackoverflow
460reputation
6kreached
1answer
2questions
Github Skills (20)
c-language10
stable10
python10
llvm10
configuration-management10
machine-learning10
build-system10
mlr10
ml10
mle10
tensorflow10
bazel10
spmd10
xla10
jax10
Programming languages (8)
Github contributions (5)
llvm/llvm-project
Nov 2023 - Mar 2025
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Automation Engineer / Build & Release Engineer
Automation Engineer / Build & Release EngineerContributions:3 reviews, 8 PRs, 8 pushes in 1 year 3 months
Contributions summary:Bart primarily focused on fixing Bazel build issues within the MLIR component of the LLVM project. Their commits addressed dependencies, configuration problems, and build failures related to code changes and upstream updates. These fixes primarily involved modifications to `BUILD.bazel` files, ensuring the successful compilation and linking of various MLIR modules and related test infrastructure. The contributions aimed to maintain build integrity and support the ongoing development and integration of MLIR features within the LLVM ecosystem.
compilerstechnologiesclangsubmittoolchain
jax-ml/jax
May 2022 - May 2022
Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
Role in this project:Backend Developer & ML Engineer
Backend Developer & ML EngineerContributions:1 commit, 4 comments, 2 issues in 1 day
Contributions summary:Bart primarily contributed to the JAX library, focusing on enhancements and modifications related to the Shardy dialect, which is part of the OpenXLA project for SPMD partitioning. Their work included implementing sharding-related features and supporting the lowering process through Shardy, specifically for with_sharding_constraint. They also addressed bugs and improved test coverage related to the new Shardy features. Their work included adding support for memories and making JAX extract the mesh from AUTO in/out sharding.
pytorchpythonjitautomatic-differentiationgpu
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.