Aparna Chandramowlishwaran

Chandramowlishwaran’s research focuses on scientific machine learning. Her group develops neural operators, transformer-based surrogate models, and foundation models that learn the structure of physical systems across conditions, scales, and domains. A growing research direction in her lab explores the mechanistic interpretability of these models, seeking to understand the representations and computations they learn and to determine whether their predictions remain faithful to the underlying physics. Drawing on expertise in high-performance computing, her group also develops efficient algorithms and implementations that enable these models to operate at the scale and speed required for practical scientific applications.