Modeled a real-world A/B test to measure homepage recommender impact on engagement. Applied causal inference, pre-trend checks, and mixed-effects modeling to identify lift drivers and user segments with the highest ROI potential.

A simulated case study for a fictional streaming platform demonstrating how product data scientists decompose core metrics and apply Bayesian causal inference to quantify the impact of subscription plans on user churn.

Ran a simulated A/B test to evaluate engagement lift from a new product feature. Used GEE and XGBoost to estimate treatment effects, identify key drivers, and validate findings with sensitivity analyses.

A collection of tools and models developed while leading federal emergency responses and public health initiatives, applying machine learning, statistical modeling, and data visualization to complex, real-world healthcare data.