I am an Assistant Professor (Praticien Hospitalier Universitaire) in Medical Informatics at Université Paris Cité, and a methodologist for the clinical data warehouse within the Medical Informatics Department at Georges Pompidou Hospital (AP-HP).

I began my scientific journey with a dual curriculum in science and medicine. My master’s thesis in computational neuroscience sparked a deep interest in machine learning, which led me to specialize in public health and medical informatics, a research-oriented specialty. During my medical residency, I developed a solid scientific foundation at the intersection of epidemiology, biostatistics, and medical informatics, contributing in particular to clinical natural language processing research.

My PhD, supervised by Francis Bach and Anita Burgun, focused “On the Design and Learning of Invariant Structures in Clinical Time Series Models”. In this work, I developed machine learning models at the interface of statistical learning—particularly Gaussian Processes—and causal structure learning.

I currently conduct my research at HeKA, focusing on the development of machine learning models for Electronic Health Records (EHRs) within the framework of Learning Health Systems. My work aims to support both clinical decision-making and biomedical research. It emphasizes the modeling of multimodal data—textual and structured—while accounting for longitudinal dynamics, uncertainty, and interpretability of predictions.

I am the recipient of the ANR-funded project LLM4ALL, in collaboration with LORIA (Christophe Cerisara and Gaël Guibon), which includes a work package on the application of large language models to medical triage.