Maternal-Fetal Risk System
Clinical vitals and CTG data brought together in a two-tier AI risk assessment dashboard.
Built with a 2-person team.
The Problem
Assessing maternal health risk and fetal distress requires working with clinical vitals and Cardiotocogram data. This project brings both into a structured dashboard and assessment API.
What I Built
A two-tier dashboard and API comparing Random Forest and XGBoost models optimized for high-risk recall.
A React/Vite frontend with role-based access, longitudinal vitals visualization, and real-time fetal assessment.
A FastAPI service with graceful fallback to clinical heuristics when the ML backend is offline.
Feature importance and permutation analysis for interpretability, plus demographic parity and equal opportunity fairness audits.
Outcome
Delivered an integrated risk assessment platform with model comparison, interpretability, fairness audits, and an offline-backend fallback.
Tech Stack