RESEARCH:

Radha Mastandrea’s research leverages physics-aware machine learning to improve model-agnostic searches for new physics. As a PhD student, Radha led a number of projects on proof-of-concept machine learning (ML)-based searches for new physics in collider data. Her PhD research projects have spanned a variety of signal signatures, such as new particles, hidden sectors, and generic parameterized effective field theories. In her past studies, Radha showed that when a search is built upon an ML architecture that is customized to the expected physics or symmetries of the signal, what results is a highly powerful and interpretable model agnostic strategy. As an Eric and Wendy Schmidt AI in Science Fellow, Radha will research and explicitly define why these symmetry-upholding ML methods have worked so well for tasks in particle physics, and she hopes to extrapolate these advances to more general classification tasks for neural networks.

BIO:

Radha Mastandrea completed her PhD at Berkeley in 2025, studying Physics with a Designated Emphasis in Computational and Data Science and Engineering. As a member of Ben Nachman’s group for Machine Learning and Fundamental Physics, she developed machine learning based methods for model-agnostic searches for new physics in particle collider data. Before her PhD, Radha received a Masters in Physics from the University of Cambridge in 2021 as a Marshall Scholar; she received her BS in Physics from MIT in 2019.