Adarsh Subbaswamy, PhD, is an Assistant Professor of Practice, Sciences, and Health Outcomes Research.
Dr. Adarsh Subbaswamy is an artificial intelligence (AI) and machine learning (ML) researcher with extensive experience applying advanced computational methods to healthcare challenges. His work focuses on developing reliable, safe, and generalizable AI algorithms for use in regulatory science and clinical decision-making. Dr. Subbaswamy brings expertise at the intersection of computer science, public health, and regulatory science, with a particular emphasis on model reliability, bias mitigation, and causal inference in health-related applications.
He previously served as a research scientist at the U.S. Food and Drug Administration (FDA), where he contributed to developing standards for evaluating AI/ML-based medical devices. His work has been published in leading venues, including npj Digital Medicine, NEJM, and Biostatistics, and he has been recognized for his contributions with awards such as the FDA OSEL Toolbox Award.
Research Interests
- Causal inference and robustness in AI/ML models
- Safe and generalizable machine learning in healthcare
- Bias detection and mitigation in predictive algorithms
- Evaluation of synthetic data and model reliability
Education and Training
- PhD, Computer Science, Johns Hopkins University
- MSE, Computer Science, Johns Hopkins University
- BS, Computer Science and Mathematics, summa cum laude, Vanderbilt University
Professional Experience
- Assistant Professor, University of Maryland School of Pharmacy (2025–Present)
- Research Scientist, FDA Center for Devices and Radiological Health (2023–2025)
- Research Fellow, FDA Center for Devices and Radiological Health (2022–2023)
Select Publications
- Subbaswamy A. et al. (2024). A data-driven framework for identifying patient subgroups on which an AI/ML model may underperform. npj Digital Medicine.
- Finlayson SG*, Subbaswamy A.* et al. (2021). The clinician and dataset shift in AI. The New England Journal of Medicine.
- Subbaswamy A., Saria S. (2020). Dataset shift, causality, and shift-stable models in health AI. Biostatistics.
- Zamzmi G., Subbaswamy A. et al. (2025). Scorecard for synthetic medical data evaluation. Communications Engineering.
- Full list available on Google Scholar.
Academic and Teaching Roles
- Instructor, AI Methodology I, University of Maryland School of Pharmacy
- Guest Lecturer at Columbia University, Johns Hopkins, UCSF, and others
- Course Developer and Co-Instructor, Summer of Machine Learning at Skoltech
- Mentor to postdocs, graduate, and undergraduate students in AI and health
Grants
- Leveraging Premarket Data to Inform Least Burdensome Performance Monitoring of AI-enabled Medical Devices — FDA (2025–Present)
- Assessing Robustness of Clinical ML Models to Context Changes — FDA (2020–2023)
Awards and Honors
- FDA Office of Science and Engineering Laboratories Toolbox Award (2023)
- Johns Hopkins CERSI Scholar in Regulatory Science (2018)
- Tau Beta Pi Engineering Honor Society
- Wilson L. and Nellie Pyle Miser Award, Vanderbilt University
Leadership and Service
- U.S. National Committee Rep, International Electrotechnical Commission (IEC)
- Program Chair, Machine Learning for Health (ML4H) Symposium
- Track Chair, Conference on Health, Inference, and Learning (CHIL)
- Reviewer for NEJM AI, Nature Machine Intelligence, npj Digital Medicine, among others