Jim Fan

Jim Fan leads robotics and physical AI research at NVIDIA, where he works on foundation models for general-purpose autonomous agents, including Project GR00T for humanoid robotics. His earlier work includes MineDojo, a Minecraft-based benchmark for open-ended agents, and research on robot foundation models and agents that learn to operate computers.

Career

Cosigned byLoading...

CosignedSign in to view

Elsewhere

NVIDIA

Dec 2021 — Present

Director and Distinguished Research Scientist

Conducting bleeding edge research on foundation models for general-purpose autonomous agents.

Columbia University

Products & media

voyager.minedojo.org

The Physical Turing Test: Solving General Purpose Robotics

A talk on evaluating general-purpose robots and the challenges of building embodied AI agents.

Media appearance

ideas.imbue.com

Foundation Models for Embodied Agents, Scaling Data, and Why Prompt Engineering Will Become Irrelevant

A conversation about embodied-agent foundation models, scaling data, and MineDojo.

Media appearance

blogs.nvidia.com

A Mine-Blowing Breakthrough: Open-Ended AI Agent Voyager Autonomously Plays ‘Minecraft’

An NVIDIA Blog feature on Voyager and how language models can enable an agent to play Minecraft autonomously.

Writing

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Jim Fan

NVIDIA

Columbia University

Products & media

Voyager: an agent that learns Minecraft

Eureka: reward design for robotics

MineDojo

The Physical Turing Test: Solving General Purpose Robotics

Foundation Models for Embodied Agents, Scaling Data, and Why Prompt Engineering Will Become Irrelevant

A Mine-Blowing Breakthrough: Open-Ended AI Agent Voyager Autonomously Plays ‘Minecraft’

Curated Lists

List Inclusions