Data Scientist - Population Health

About This Role

Tile Health’s population health analytics suite empowers care teams to intervene before costly health events occur. The Data Scientist for Population Health will develop and validate predictive models for outcomes such as hospital readmission, ED utilization, and chronic disease progression. You will work with rich longitudinal datasets spanning claims, EHR, and social determinants to generate insights that directly impact patient care.

What You'll Do

• Build and validate predictive models for clinical outcomes including 30-day readmission, ED recidivism, and disease progression risk
• Conduct exploratory analyses on multi-source datasets (EHR, claims, SDOH) to identify novel risk factors and intervention targets
• Design and execute A/B tests and quasi-experimental analyses to measure the causal impact of clinical programs
• Create reproducible analysis pipelines using Python and SQL with version-controlled notebooks
• Partner with clinical operations to translate model outputs into actionable care management workflows
• Communicate findings through clear visualizations and written reports for both technical and clinical audiences

What We're Looking For

• Master’s or PhD in Biostatistics, Epidemiology, Data Science, or a related quantitative field
• 3+ years of applied data science experience with healthcare datasets (claims, EHR, or registry data)
• Proficiency in Python (scikit-learn, pandas, statsmodels) and SQL for data analysis
• Experience with survival analysis, logistic regression, and gradient-boosted tree methods
• Understanding of healthcare-specific evaluation metrics including calibration, net benefit, and clinical utility curves

Nice to Have

• Experience with causal inference methods (difference-in-differences, instrumental variables, propensity score matching)
• Familiarity with OMOP Common Data Model or PCORnet data standards
• Published research in health services research, outcomes research, or clinical epidemiology
• Experience building models that satisfy fairness constraints across racial and socioeconomic subgroups