AI Security Operations

What you will be able to do

• You will be able to map the AI and machine learning attack surface using MITRE ATLAS (Adversarial Threat Landscape for Artificial Intelligence Systems) and the OWASP (Open Worldwide Application Security Project) Top 10 for LLM (Large Language Model) Applications.
• You will be able to instrument a production model with telemetry, drift detection, and guardrails, and write detections that fire on abuse instead of on normal traffic.
• You will be able to triage AI alerts and tell a real prompt injection, jailbreak, or model-abuse attempt apart from a benign edge case.
• You will be able to run an AI incident from detection through containment, investigation, and recovery using a repeatable playbook.
• You will be able to harden the machine learning pipeline against data poisoning and tampered model artifacts across the training and deployment lifecycle.
• You will be able to assess third-party model and dataset risk and maintain an AI Bill of Materials (AI-BOM) that tracks model and data provenance.
• You will be able to connect day-to-day AI security operations to the NIST AI RMF and ISO/IEC 42001, and produce the evidence that governance and auditors expect.

What is inside

4 modules, 8 lessons. Each module ends in a graded quiz and most carry an assignment.

1. 01

   Monitoring AI in Production

   You cannot defend a model you cannot see. This module maps the AI and machine learning attack surface with MITRE ATLAS and the OWASP Top 10 for LLM Applications, then shows how to instrument live models with telemetry, drift detection, and guardrails. By the end you can stand up monitoring that turns raw model behavior into signals a security team can actually act on.

   2 lessons · 5 quiz questions

2. 02

   Detecting and Responding to AI Incidents

   Monitoring only matters if someone acts on the alert. This module teaches you to triage AI-specific signals, separating real prompt injection, jailbreaks, and model abuse from noisy false positives, then to run the incident with a repeatable playbook. You will work through containment, investigation, and recovery for attacks that classic incident response was never designed to handle.

   2 lessons · 5 quiz questions · assignment

3. 03

   Securing the ML Pipeline and Supply Chain

   Most AI risk is built in long before a model reaches production. This module moves upstream into MLSecOps: you learn to protect training data and model artifacts from poisoning and tampering, and to harden the pipeline that trains, packages, and deploys them. You also learn to manage the AI supply chain, from untrusted model files to third-party providers, using provenance and an AI Bill of Materials (AI-BOM).

   2 lessons · 5 quiz questions

4. 04

   Connecting Operations to AI Governance

   Frontline AI security only holds up when it feeds the governance program above it. This module connects your day-to-day operations to the NIST AI RMF and ISO/IEC 42001, showing how monitoring data, incidents, and pipeline controls become governance evidence, and how to report AI security metrics and continuous assurance in language that leadership and auditors trust.

   2 lessons · 5 quiz questions · assignment