Jason Poulos

Jason Poulos is a data scientist and researcher currently working at the intersection of machine learning, causal inference, and health policy. He holds a Ph.D. in Political Science from the University of California, Berkeley, with a designated emphasis in Computational Science and Engineering. His academic journey includes postdoctoral positions at Harvard Medical School's Department of Health Care Policy and Brigham and Women's Hospital.

Dr. Poulos's research focuses on developing and applying machine learning methods to address challenges in complex, high-dimensional longitudinal health data. His work spans several interconnected areas including extending causal inference methods for multi-valued treatments, developing frameworks for adversarial machine learning, applying neural networks for health outcome prediction, and evaluating deep learning approaches for missing data imputation in survey data. His recent publications demonstrate expertise in targeted learning methodologies, particularly in observational studies with multi-level treatments, with applications in pharmacoepidemiology.

His research output shows a consistent pattern of interdisciplinary work bridging political science, statistics, and health policy. The majority of his recent publications (2021-2024) appear in high-impact journals across biostatistics, political methodology, and machine learning domains, reflecting his ability to contribute to multiple academic communities. His work often employs advanced statistical techniques to address real-world policy questions, particularly in healthcare settings.

As evidenced by his publications and research trajectory, Dr. Poulos has established himself as a methodological innovator in causal inference with multi-valued treatments, with particular expertise in applying these methods to evaluate the safety and effectiveness of antipsychotic medications. His work combines deep methodological rigor with practical policy applications.