A predictive maintenance model for estimating machine remaining useful life (RUL) with uncertainty. Instead of predicting a single value, the model outputs a range of plausible outcomes, enabling more informed maintenance decisions.

End-to-end predictive maintenance ML system with FastAPI inference, MLflow tracking, and Docker deployment

RUL prediction on NASA C-MAPSS with Random Forest — MAE ≈ 12 cycles

An autonomous civic tech platform for decentralized hyperlocal infrastructure auditing. Features peer-to-peer hazard reporting, Gemini AI-driven predictive diagnostics, and a Women's SafeGuard Hub for emergency SOS coordination.

Spring Boot backend that ingests industrial IoT sensor telemetry, applies threshold-based anomaly detection across temperature/vibration/pressure metrics, and exposes equipment health via REST APIs.