PNNL

Soumya is a data scientist in the Auto Learning and Reasoning team, Physical & Computational Science Directorate. She was previously a post-doc research associate in the Analytics and Learning team within the Optimization and Control group. Her work draws from various disciplines including optimal control theory, deep learning, reinforcement learning, and probabilistic programming for improved system design, modeling, and control. Her current interests lie in physics-informed formulations that bridge the gap between traditional deep learning methods and classical system and control analysis paradigms to realize more powerful, reliable, robust control-oriented system representations. She received her PhD in aeronautics and astronautics from the University of Washington in 2019. Her work centered on building data-driven algorithms for navigation, estimation, and tracking for autonomous vehicles. At Pacific Northwest National Laboratory, she has been involved in modeling, optimization, and uncertainty quantification for building energy systems, physics-informed analysis of generic deep learning models, and parking prediction models for dynamic curb allocation.

• PhD, Aeronautics and Astronautics – Robotics and Data-driven Control, University of Washington, Seattle, WA (2019)
• MS, Aeronautics and Astronautics – Dynamics and Control, University of Washington, Seattle, WA (2013)
• BEngg, Information Science and Engineering, P.E.S. Institute of Technology, Bangalore, India (2007)

• IEEE Control Systems Society, Women in Engineering
• Organizer:
  • AMLIES Workshop at ACM e-Energy Conference (2021)
  • Internal PNNL Physics-informed Machine Learning Workshop (2021)
• Reviewer:
  • IEEE Transactions on Aerospace and Electronic Systems, American Control Conference, Learning for Dynamics and Control, Climate Change AI grants

• Varanasi-endowed Fellowship (2013)