Selected work across AI governance, national health data, and applied machine learning. Each project links to the original publication or framework.
Developed clinical NLP models to identify diabetes-related complications in unstructured Hebrew medical records, inside the privacy and security constraints of a national health-data environment.
Applied machine learning to longitudinal HbA1c records from the Israel National Diabetes Registry and examined their association with all-cause mortality.
Built and evaluated machine-learning models for estimating the risk of an unplanned Cesarean delivery from routinely collected clinical data.
Used records from pediatric clinics across Israel to standardize the national developmental milestone scale and establish population reference values.
A multi-year longitudinal study of the association between early postpartum depressive symptoms and later child development outcomes.