Tess Smidt

Tess Smidt is an Associate Professor of Electrical Engineering and Computer Science at MIT. Tess earned her SB in Physics from MIT in 2012 and her PhD in Physics from the University of California, Berkeley in 2018. Her research focuses on developing machine learning methods that incorporate physical and geometric structure to model and design molecules, materials, and other complex physical systems. She is a 2025 AI2050 Early Career Fellow of Schmidt Sciences and a recipient of the DOE Early Career Award and AFOSR Young Investigator Research Program (YIP) Award. Before joining MIT, she was the Alvarez Postdoctoral Fellow in Computing Sciences at Lawrence Berkeley National Laboratory, a Software Engineering Intern on the Google Accelerated Sciences, and co-developed early Euclidean symmetry-equivariant neural networks for learning from 3D geometric data.
AI2050 Project

Smidt’s project builds AI systems that understand and respect the deep structure of the physical world, its symmetries, hierarchies, and complex dynamics. By designing models that are aware of physical laws and operate across multiple scales, from atoms to galaxies, they aim to create tools that not only predict physical behavior but help scientists explore, control, and design it. These symmetry-aware AI models will enable breakthroughs in materials discovery, fluid dynamics, and cosmology by making abstract scientific principles actionable, interpretable, and generative, paving the way for AI to become a true partner in scientific reasoning and innovation.