Author: Google Cloud Tech - Bewertung: 1x - Views:0
Multi-agent systems are increasingly vulnerable to prompt injections, data leaks, and denial-of-service attacks when their endpoints are left exposed. In this session, developers will learn how to implement a comprehensive defense-in-depth strategy to secure AI agents using Google Cloud’s native security tools.
*Join us and learn:*
* How to implement Sensitive Data Protection (SDP) to redact PII before it reaches agents.
* How to integrate Model Armor to intercept and sanitize user prompts against malicious injections.
* Learn the workflow for running live Red Team scenarios to stress-test defenses.
* How to manage security guardrails at scale using "Policy as Code" with Terraform.
* How to apply defense-in-depth principles to protect multi-agent architecture.
* How to ensure consistent security posture across development and production environments.
*Resources:*
* Codelab: Securing a Multi-Agent System → https://goo.gle/3T9Z2FK
* Google Cloud Sensitive Data Protection documentation → https://goo.gle/4h6gu8b
* Model Armor overview → https://goo.gle/4hswp0m
* Google Cloud AI security best practices → https://goo.gle/4pDltPS
Ready to build? Try this in Google Cloud Codelabs here. → https://goo.gle/4vGkaAX
#ModelArmor #CloudSecurity
Speakers: Aron Eidelman, Leonid Yankulin
Products Mentioned: Gemini, Model Armor, Cloud Security