NVIDIA & Marlowe: Training Robots with World Foundation Models

Abstract

This talk explores the new wave of robotics driven by world foundation models. We present our family of world models and describe how they are integrated into our ecosystem to scale robot training, simulation, and generalization across diverse tasks and environments.

Special Speakers | Akul Santhosh & Aiden Chang

Akul Santhosh is a Solutions Architect at NVIDIA, working on the Robotics team. He collaborates with researchers and developers working on simulation technogies and world foundation models. Before joining NVIDIA, he earned a master’s degree in computer science from New York University with a focus on Robotics and Simulations.

Aiden Chang is a Solutions Architect at NVIDIA, focusing on enterprise applications of generative AI, robotics, and reasoning systems. He earned his M.S. in Computer Science from the University of Southern California. Outside of work, he enjoys skiing, aviation, and building robots.