Predictive Maintenance Platform for a Manufacturing Client
Gen AI CM built a predictive maintenance layer for a manufacturing firm’s plant-level IoT data, targeting downtime prevention and optimal inventory planning. The system monitored 40+ machines in real time and triggered alerts pre-failure.
Technology / Tools / Frameworks Used:
- AWS SageMaker (Random Forest, XGBoost)
- Time-series anomaly detection with Facebook Prophet
- Kafka + Spark Streaming for real-time telemetry
- Edge computing deployment with S3 + Greengrass
- Dashboard hosted via Grafana with Okta-based access control
Client Value Delivered:
- Prevented over ₹1.6 Cr in unplanned downtime
- Predicted failures with 93% precision
- Enabled just-in-time spare part procurement
- Optimized maintenance workflows across 3 factories
