Samuel Schoenholz
Founding Team at Thinking Machines Lab
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSamuel Schoenholz is a machine learning engineer and researcher with 11 years of experience bridging statistical physics and practical ML, now on the founding team of Thinking Machines Lab in Mountain View. He spent several years at Google and OpenAI working on research-driven engineering—applying theoretical tools from disordered systems to advance ML—and has a PhD in Physics from the University of Pennsylvania. Samuel is a prolific open-source contributor to flagship projects in the JAX/TensorFlow ecosystem, notably improving batching, convolution/pooling rules, and portability in tensorflow/probability, JAX, and Neural Tangents. His work combines deep mathematical insight with hands-on backend engineering to make advanced algorithms robust and cloud-friendly, a blend informed by prior academic work on glassy materials and statistical mechanics.
11 years of coding experience
14 years of employment as a software developer
High School, High School at International School of Geneva - La ChataigneraieBachelor of Arts (B.A.) Physics Mathematics, Bachelor of Arts (B.A.) Physics Mathematics at Swarthmore CollegeDoctor of Philosophy (PhD) Physics, Doctor of Philosophy (PhD) Physics at University of Pennsylvania
English, French
Github Skills (22)
batching10
molecular-simulation10
python10
artificial-neural-networks10
machine-learning10
numpy10
automatic-differentiation10
latex10
molecular-dynamics-simulation10
tensorflow10
nonlinear-dynamics10
neural-network10
batch10
jax10
ai9
Programming languages (4)
Github contributions (5)
jax-md/jax-md
May 2019 - Dec 2022
Differentiable, Hardware Accelerated, Molecular Dynamics
Role in this project:
Back-end Developer
Back-end DeveloperContributions:6 releases, 30 reviews, 687 commits in 3 years 7 months
Contributions summary:Samuel's contributions primarily involve small changes and bug fixes to existing code. The edits include wording changes to documentation, adjustments to notebook settings, and the correction of a few minor errors. Furthermore, the user's work focuses on refactoring and adapting simulation settings.
differentiablemolecular-dynamicsautomatic-differentiationacceleratedsimulation
google/neural-tangents
Apr 2019 - Apr 2022
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:54 commits, 5 PRs, 104 pushes in 3 years
Contributions summary:Samuel primarily contributed to improving the installation instructions and ensuring the notebooks run correctly in cloud environments. They added code to remove Google-specific imports, simplifying the notebooks' portability. Furthermore, they refactored kernel functions throughout the library, allowing for more flexible specification of data of interest. The user demonstrated expertise in the Neural Tangents library by modifying and optimizing its core functionalities.
kernelpythonbayesian-inferencebayesian-networksgaussian-processes
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.