A selection of applied research and strategic frameworks. This work demonstrates the connection between high-level governance and the technical execution of building AI models on sensitive, real-world data.
Architected and guided the data science strategy for TIMNA, Israel's national big data platform for medical research. TIMNA securely integrates clinical and genomic datasets across the country's entire healthcare ecosystem, enabling advanced analytics while meeting strict sovereign compliance, national privacy, and security controls.
Led the Ministry of Health system for mapping COVID-19 infection chains — the national contact-tracing data backbone behind Israel's epidemiological investigations. Under extreme wartime-pace pressure, the system linked case data, exposure events, and investigation workflows at national scale, turning fragmented reports into actionable chains that guided isolation policy and outbreak containment in real time.
Designed and executed clinical Natural Language Processing (NLP) models to extract diabetes-related complications from unstructured Hebrew Electronic Medical Records (EMR). This project showcases the ability to deploy text-processing pipelines in highly restricted environments, turning messy, multilingual clinical narratives into structured, actionable operational data.
Developed and applied machine learning models to analyze longitudinal HbA1c trends and all-cause mortality across the Israel National Diabetes Registry. This work demonstrates how to run predictive modeling on large-scale national databases, connecting historical patient data with risk stratification to guide preventative medicine.
Built predictive machine learning models to identify risks for unplanned Cesarean deliveries. This research translates complex predictive modeling into clinical decision support, illustrating how raw predictive algorithms can be operationalized to assist clinical workflows and drive immediate medical value under strict safety bounds.
Standardized the national developmental milestone scale using big data from pediatric clinics in Israel. This research established a data-driven baseline for child development, proving how nationwide databases can be securely integrated to drive national healthcare policy and support clinical decision-making.
A multi-year longitudinal study evaluating the association of early postpartum depressive symptoms with long-term infant development. This project demonstrates how long-running registry data and outcomes-tracking frameworks can be operationalized to influence evidence-based public health interventions.