Jonathan Tompson - Principle Research Scientist (director) at Meta

Jonathan Tompson

Principle Research Scientist (director) at Meta

San Francisco, California, United States

Join Prog.AI to see contacts

Summary

👤

Senior

🎓

Top School

Jonathan Tompson is a Principal Research Scientist and director based in San Francisco with 13 years of experience leading robotics and machine learning efforts across industry and academia. He has held senior research leadership roles at Google DeepMind and now leads Robotics ML at Meta Reality Labs, blending hands-on research, engineering, and team management. His background spans a PhD from NYU Courant and strong engineering foundations from Harvard and Columbia, with experience building vision-based tracking startups and shipping production software. Jonathan is an active open-source contributor to high-profile projects like FluidNet and Bullet Physics, where he implemented CUDA-backed neural modules and improved core simulation and shared-memory systems. He brings a rare combination of low-level systems expertise (CUDA, physics engines, build systems) and applied ML for robotics, often bridging research prototypes to scalable production code. Colleagues describe him as a pragmatic researcher who routinely turns complex numerical and simulation challenges into reliable, deployable systems.

13 years of coding experience

14 years of employment as a software developer

Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at New York University

B.S EE & CS Electrical Engineering and Computer Science, B.S EE & CS Electrical Engineering and Computer Science at Harvard University

M.S EE Electrical Engineering, M.S EE Electrical Engineering at Columbia Engineering

High School Diploma, High School Diploma at The Friends' School, Hobart, Tasmania

Github Skills (37)

simulation10

pytorch10

c-language10

operation10

simulator10

tensorrt10

simulations10

machine-learning10

kernel10

lua10

fluid-simulation10

deep-learning10

tensorflow10

physics10

cuda10

Programming languages (8)

C++CCMakeLuaJupyter NotebookMATLABPythonCuda

Github contributions (5)

google/FluidNet

Nov 2016 - Nov 2022

Accelerating Eulerian Fluid Simulation With Convolutional Networks

Role in this project:

Full-stack Developer

Contributions:32 commits, 11 PRs, 125 pushes in 6 years

Contributions summary:Jonathan's contributions focused on modifying and refactoring the codebase related to Eulerian fluid simulation. They updated model parameters and fixed issues in the README.md file. Furthermore, the user modified the build structure to integrate with Google's internal blaze build system. They also added cusparse support and included the skeleton code for PCG implementation.

fluid-simulation

torch/cunn

Sep 2013 - Jun 2016

Role in this project:

Back-end Developer

Contributions:15 commits, 5 PRs, 25 comments in 2 years 9 months

Contributions summary:Jonathan primarily contributed CUDA implementations for various mathematical and neural network operations within the cunn library, which is a CUDA backend for the Torch deep learning framework. They added modules like Abs, Exp, SoftPlus, Min, and DistKLDivCriterion, along with their corresponding tests. The user also implemented the ReLU6 activation function and added a volumetric replication padding operation for spatial data processing.

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.

Request Free Trial