Built a predictive early-warning system that identifies at-risk users before they disengage, using only early behavioral data. Achieves 84%+ accuracy with built-in explainability so stakeholders understand why someone is flagged, not just that they are. Directly applicable to SaaS platforms, e-learning products, and subscription businesses looking to reduce churn.