Brian Patton
Software Engineer at Google
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolBrian Patton is a seasoned software engineer with 15 years of experience building production-grade systems at Google from frontend analytics and site reliability to probabilistic machine learning. Based in Zurich, he combines deep numerical and compiler-level expertise—contributing to high-impact open-source projects like JAX, XLA and TensorFlow Probability—with pragmatic engineering that modernizes legacy ML code and improves numerical stability. His work spans C++, Python, and ML toolchains, including implementing statistical distributions, kernel refactors, and GPU-focused performance fixes. Known for smoothing build and platform issues across macOS, GPU/TPU and cloud environments, he bridges research-grade algorithms and robust, maintainable code. Less obvious: he has repeatedly navigated both low-level systems and probabilistic modeling, making him effective at shipping complex ML infrastructure that is both fast and numerically reliable.
16 years of coding experience
BA, Computer Science, BA, Computer Science at Boston University
Github Skills (30)
algorithm10
c-language10
subsampling10
numerical10
python10
machine-learning10
sampling10
distributions10
numpy10
numerics10
tensorflow10
numerical-methods10
sampler10
ai10
stability10
Programming languages (9)
Github contributions (5)
tensorflow/probability
Jul 2018 - Jan 2023
Probabilistic reasoning and statistical analysis in TensorFlow
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:5 releases, 18 reviews, 641 commits in 4 years 6 months
Contributions summary:Brian contributed to the development of the VonMisesFisher distribution and related sampling methods, adding functionalities such as sampling and log probability computations. They implemented numerical improvements by modifying and debugging the Bessel function implementations used in the pdf, and fixed multiple bugs and errors in the core mathematical formulas for statistical distributions. They demonstrated expertise in probability distributions and numerical stability within a Tensorflow Probability context.
probabilistic-reasoningtensorflowbayesian-methodsdeep-learningmachine-learning
jax-ml/jax
Jun 2019 - Mar 2026
Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
Role in this project:Back-end Developer
Back-end DeveloperContributions:1 review, 6 PRs, 44 comments in 6 years 10 months
Contributions summary:Brian primarily focused on improving the functionality and stability of the JAX library. Their contributions include making core components weakref-friendly, ensuring correct data type handling, and addressing potential exceptions. The user also optimized performance by unrolling the `scan` operation in `jnp.searchsorted` for GPU utilization. They also fixed a bug and ensured effect indices were updated.
gpujittpujax
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.