Model Armor: Securing AI Deployments
This 2-3 hour Cloud Sparks session is a short, interactive course on protecting AI applications with Model Armor.
The course explains how to use Model Armor to protect large language model (LLM) applications from threats such as malicious URLs, prompt injection, jailbreaking, sensitive data leaks and improper output handling. You will learn about its architecture, how its features map to the OWASP top 10 LLM risks it addresses, and how it manages prompts and responses.
Practical skills include defining floor settings, configuring templates and enabling the different detection types, setting up the Model Armor API, and exploring audit logs to find details about flagged violations. The course includes a lab.
This course is aimed at security engineers, AI/ML developers and cloud architects. It is recommended that attendees have working knowledge of APIs, the Google Cloud CLI and console, and cloud security fundamentals.