Jiatao Gu

Jiatong Gu is an Assistant Professor in the Department of Computer and Information Science at the University of Pennsylvania, where he is affiliated with the GRASP Lab and the ASSET Center. He leads the Generative Multimodal Learning and Reasoning (GMLR) Lab, which was established in 2024 and focuses on building foundational models for generalist agents. His research aims to develop generative world models that enable long-horizon prediction, memory, planning, and decision-making across complex environments. Core research areas include multimodal generative models spanning images, video, and 3D/4D representations; world models for sequential reasoning; vision-language and vision-language-action models for task decomposition; and agent learning with applications in robotics and embodied AI. Prior to returning to academia, he spent several years as a Staff Research Scientist at Meta FAIR and Apple, where he worked on large-scale multimodal learning systems. Since founding the GMLR Lab, he has recruited PhD students and researchers across multiple levels and is actively building a research group at the intersection of multimodal generation, reasoning, and intelligent agents.