Somya Chatterjee

About

Somya is an applied scientist 2 at Microsoft’s AI Development and Acceleration Program who develops LLM-powered and Agent-based systems for powering Microsoft’s Copilot solutions. Her research work focuses on driving performance and efficiency of generative AI solutions. Somya completed her PhD in Computer Science at the University of Minnesota – Twin Cities, received MS in Statistics from the School of Statistics at the University of Minnesota – Twin Cities and BS in Statistics (Honors) from the University of Delhi. Her PhD research focuses on generative modeling and explainable AI for deep learning applications in physical and geosciences research. Somya also worked in the Research for Industry team at Microsoft Research and collaborated with the Global Soil Health Program to develop causal structure learning and causally adaptive, large-scale deep learning models. Her research addresses the challenges in Precision Agriculture, Wildfire Management, Hydrology, and Molecular Science. Her research has been presented at several highly selective, peer-reviewed ML conferences like NeurIPS and SDM. Some of her research also won the Doctoral Forum Award at SIAM Internation Conference on Data Mining, has received media coverage (Microsoft, MetroTransit, NOAA NWS news), has been highlighted as spotlight talks at NeurIPS, and has been awarded university research fellowship awards (e.g., Doctoral Dissertation Research Fellowship).