Jim Fan
Director & Distinguished Research Scientist at NVIDIA
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJim Fan is a Director and Distinguished Research Scientist at NVIDIA leading efforts to bring Physical AI to reality through foundation models for general-purpose robotics. With a Stanford PhD and 12 years of experience spanning OpenAI, MILA, Google, and multiple NVIDIA research roles, he builds multimodal generalist agents that bridge simulation and the real world. He co-founded and co-leads the GEAR team and spearheads Project GR00T, focusing on agents that learn to act across virtual and physical environments. His work has won top honors such as a NeurIPS Best Paper for MineDojo and includes hands-on contributions to infrastructure like GPU-enabled ML tooling and headless-simulation Docker stacks for iGibson. Based in Palo Alto, he combines deep academic rigor with product-scale engineering and a track record of shipping foundational systems that power embodied intelligence. Outside core research he actively documents and shares his work (Google Scholar, GitHub, Twitter), reflecting a commitment to open science and reproducible robotics.
12 years of coding experience
9 years of employment as a software developer
Bachelor of Science (B.S.), Computer Science, GPA: 4.3/4.3, Bachelor of Science (B.S.), Computer Science, GPA: 4.3/4.3 at Columbia University in the City of New YorkDoctor of Philosophy - PhD, Computer Science, 4.0, Doctor of Philosophy - PhD, Computer Science, 4.0 at Stanford University
English, Chinese
Github Skills (16)
pytorch10
bash10
docker10
tensorflow10
gpu10
devops10
python10
azure-devops10
dockers10
gcp9
machine-learning9
cuda9
deep-learning8
3ds8
3d8
Programming languages (12)
Github contributions (5)
cs231n/gcloud
Mar 2019 - Apr 2019
Google Cloud tutorial and setup
Role in this project:
ML Engineer
ML EngineerContributions:9 commits, 10 pushes, 1 comment in 16 days
Contributions summary:Jim focused on setting up and verifying GPU usage within the Google Cloud environment. They implemented and updated a script to check for CUDA availability and display GPU information. Furthermore, the user added support for TensorFlow, including a sample MNIST model and a setup script that installed TensorFlow-GPU. These contributions suggest an emphasis on utilizing GPU resources for machine learning tasks within the Google Cloud context.
google-cloud
StanfordVL/iGibson
Apr 2020 - Aug 2020
A Simulation Environment to train Robots in Large Realistic Interactive Scenes
Role in this project:DevOps Engineer
DevOps EngineerContributions:26 commits, 2 PRs, 1 comment in 4 months
Contributions summary:Jim primarily focused on building and maintaining the project's infrastructure by creating and modifying Docker images to support headless GUI functionality. They implemented scripts to pull and push Docker images, automating the build process and facilitating the testing and deployment of the iGibson environment. The user also made changes to support environment variables to enable easier configuration.
simulationrobot-simulator3d-scenes
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.