Antonio Olivas-Martinez, PhD, MD
Postdoctoral Researcher
Antonio Olivas-Martinez, PhD, MD, is a postdoctoral researcher in Biostatistics and Epidemiology within the Department of Biostatistics, Epidemiology, and Informatics at the University of Pennsylvania’s Perelman School of Medicine. He is affiliated with the Center for Causal Inference (CCI) and the Center for Health Innovations in Reproductive and Perinatal Population Research (CHIRP). His work is mentored by Eric Tchetgen Tchetgen, Enrique Schisterman, and Ellen Caniglia.
His current research focuses on causal inference under unmeasured confounding, including proximal methods; regularized estimation for ill-posed inverse problems; and updating SART IVF prediction models using flexible statistical and supervised learning methods.
Antonio earned a PhD in Biostatistics from the University of Washington and previously trained in mathematics, medicine, and internal medicine in Mexico. His interdisciplinary background motivates his development of rigorous statistical methods that remain clinically interpretable and useful for health decisions. He also maintains research collaborations with investigators at the Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán and the University of Sonora (Universidad de Sonora) in Mexico, as well as through the Global Burden of Disease Collaborator Network.
Research Areas
Causal Inference, Novel Methods, Translational Research, Semiparametric Theory, Nonparametric and Regularized Estimation, Ill-posed Inverse Problems, Supervised Learning
Education
- PhD, Biostatistics, University of Washington 2025
- Specialty in Internal Medicine, Universidad Nacional Autónoma de México / Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán 2019
- Medical Degree, Universidad de Sonora 2014
- BMath, Mathematics, Universidad de Sonora 2008
Contact
Research Highlights
Causal Inference with Unmeasured Confounding
Overview: Development and application of proximal and sensitivity-analysis methods for settings where important confounders cannot be measured.
Regularized Estimation for Ill-Posed Inverse Problems
Overview: Source-condition analysis and finite-sample error bounds for kernel regularized adversarial estimators, together with a framework for comparing the assumptions underlying related methods.
SART IVF Prediction Models
Overview: Updating models used to predict in vitro fertilization outcomes within CHIRP using flexible statistical and supervised learning methods.