Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

AI Threat Modeling and Operational Defense

Overview

Secure the AI systems your organization is putting into production. You will threat model machine learning architectures with STRIDE-ML, inventory model and dataset dependencies in an ML-BOM, and harden inference endpoints against extraction and abuse. You will filter prompt injection with Bedrock Guardrails, scope IAM roles so a compromised agent cannot reach what it never needed, trace poisoned documents back through a knowledge base, and rate limit an API so runaway cost becomes a manageable event. The work is hands-on in AWS with Python, and it builds toward a capstone where you deploy, harden, and validate a Bedrock RAG agent. Bring working knowledge of Python, cloud fundamentals, and basic security concepts.

Syllabus

Welcome to AI Security

Set up your tooling and AWS access, and preview the AI security skills and the Bedrock capstone project ahead of you.

Build Aria: Set Up the AWS Exercise and Demo Environment
Understand Threat Modeling for Machine Learning

Learn how classic threat modeling adapts to machine learning, and use STRIDE-ML to name specific threats against each part of an AI system.

Apply Threat Modeling with STRIDE-ML

Build a structured threat model for a Bedrock RAG architecture, ranking threats by likelihood and impact to pick the control that ships first.

Understand the Machine Learning Bill of Materials

Track pre-trained models and datasets as supply chain dependencies, and see what an ML-BOM records so you can scope a vulnerability fast.

Generate an ML-BOM for a Machine Learning Project

Create an ML-BOM for a RAG system, documenting both the embedding and generation models plus the provenance gaps a managed API leaves behind.

Understand Secure Model Serving and API Management

Examine why an inference endpoint is a security boundary of its own, and the layered controls that protect a model from theft and abuse.

Audit Bedrock Client Security Configuration

Audit a Bedrock client configuration for hardcoded credentials, missing input validation, and absent logging, then scope its IAM policy correctly.

Understand AI-Specific Monitoring and Incident Response

Discover why AI systems fail behaviorally while infrastructure stays healthy, and which signals reveal drift, hallucination, and active attacks.

Configure Bedrock Logging and Incident Response Playbooks

Configure model invocation logging, then write a monitoring plan and an incident response playbook for an AI-specific event.

Understand Input Sanitization for LLM Prompts

Study how prompt injection works, including payloads hidden in retrieved documents, and where denylists and allowlists each fall short.

Filter Input and Output with Bedrock Guardrails

Design and test Bedrock Guardrails that block injection attempts on the way in and catch sensitive data on the way out.

Configure PII Detection and Data Classification

Classify documents by sensitivity and configure PII detection so a knowledge base carries only what the assistant actually needs.

Configure IAM Policies for Cloud AI Services in AWS

Write least-privilege policies and an organization-level SCP, then interpret IAM Access Analyzer findings for Bedrock resources.

Understand Least Privilege for LLM Tools

Assess agent tools by the damage they can do, and see why a human approval step belongs in system architecture rather than in a prompt.

Enforce Least Privilege with IAM Roles

Right-size the IAM roles behind a RAG agent, replacing wildcards with scoped ARNs and measuring the blast radius you remove.

Understand Data Provenance Tracking

Trace what a complete provenance record contains, and how poisoned documents reach a knowledge base and stay hidden inside it.

Detect and Investigate Knowledge Base Poisoning

Investigate a poisoned knowledge base, walking a provenance trail backward from a bad response to the upload event that caused it.

Understand Rate Limiting for Inference APIs

Compare fixed and sliding window rate limiting, and see why unbounded cost threatens an inference API as much as downtime does.

Implement a Sliding Window Rate Limiter in Python

Implement a sliding window rate limiter, analyze usage logs for abuse patterns, and set a limit you can defend from expected traffic.

Understand IAM Policies for Cloud AI Services

Read an IAM policy for a cloud AI service and identify the wildcard patterns that hand an attacker administrative control.

Project: Northstar Assist: Build and Secure an AWS Bedrock RAG AI Agent

Deploy, harden, and validate Northstar Assist, an Amazon Bedrock AgentCore RAG agent, applying threat modeling, an ML-BOM, guardrails, and logging to a system that is ready to launch.