Anomaly Detection in Hardware Systems

End-to-end pipeline for anomaly detection using ML, IoT sensors, and AWS Cloud.

Developed a complete anomaly detection pipeline to monitor hardware vibrations using 50 sensor nodes integrated with Raspberry Pi for data collection and transmission.

  • Designed and deployed REST APIs to transfer sensor data from Raspberry Pi to AWS Cloud.
  • Built and trained machine learning models (Isolation Forest, Random Forest, Autoencoder) to detect real-time anomalies with up to 93.7% accuracy.
  • Developed a real-time dashboard for visualizing sensor data and anomaly alerts.
  • Deployed the entire solution on AWS Cloud ensuring scalability, reliability, and efficient data stream handling.