Fitsum Reda

Fitsum Reda

Principal Scientist

NVIDIA

About

I’m a principal scientist at NVIDIA, where I work on Nemotron large language models and previously on Cosmos world foundation models. I’m excited about advancing large language models, multimodal understanding, and world modeling, and their convergence as a path to superintelligence that can reason, create, and learn from experience.

Previously, I co-created FILM and TryOnDiffusion at Google Research and co-founded and led the frame generation research behind NVIDIA DLSS 3.

I earned a PhD in electrical engineering from Vanderbilt University and an MS in computer vision and robotics from Heriot-Watt University.

Interests
  • Large Language Models
  • World Foundation Models
  • Multimodal Language Models
Education
  • PhD in Electrical Engineering

    Vanderbilt University

  • MS in Computer Vision and Robotics

    Heriot-Watt University

Contact