Monitoring an ML-Based Intrusion Detection System on AWS SageMaker

1. What We Are Building We are going to deploy a Random Forest classifier trained on the UNSW-NB15 dataset as a real-time network intrusion detection system (IDS). The model classifies live network flows into either Normal traffic or one of nine attack categories: Fuzzers, Analysis, Backdoors, DoS, Exploits, Generic, Reconnaissance, Shellcode, and Worms. Once deployed, we will configure SageMaker Model Monitor to detect when the model starts degrading, wire CloudWatch alarms to page the on-call