Loading
Health Tech
AI-powered maternal mortality risk prediction and clinical workflow platform that gives community health workers in Cameroon a complete digital toolkit from antenatal care through postnatal and infant monitoring.
Maternal mortality remains critical in sub-Saharan Africa, where most deaths are preventable with early detection and timely referral. The root problem is not a lack of clinical knowledge, it is a lack of tools that work at community level, offline, in the hands of CHWs who carry paper registers and make decisions without decision support.
Built a full-stack system around a SHAP-explainable XGBoost risk classifier: a FastAPI/PostgreSQL backend drives ANC visit scheduling (auto-generating the WHO 8-visit schedule), emergency referral tracking via SMS, and longitudinal risk trend monitoring with escalation alerts. A React web dashboard serves supervisors and facility management, while a Flutter offline-first mobile app and a WhatsApp gateway (Baileys) put reminders and alerts directly in CHWs' hands on basic smartphones, the whole stack deployed single-server via Docker Compose behind Nginx with TLS.
Delivered a production-ready, single-server deployment covering risk prediction, ANC/PNC visit tracking, infant growth monitoring (WHO growth standards), and a district-level facility dashboard with Ministry of Health report export, built around the Three Delays Framework to cut time-to-referral for maternal emergencies.
