AI Security for Kubernetes: Guardrails, Red Team, and Remediate
summary
F5 runs its AI security stack in Kubernetes and routes model traffic through the F5 NGINX Ingress Controller and Gateway Fabric for inspection. Guardrails include keyword, regex, and genAI scanners that detect, redact, block and log prompt injection, jailbreaks or sensitive data. Red teaming runs agentic pentests against deployed or preprod models to surface vulnerabilities. A newly announced Remediate integrates red team findings back into testing so discovered exploits feed automated retests and validation.
AI-written summary · about
related talks
39mHow to Reconcile Security and Agility for the Good of Your Applications
16mKeynote: Defending AI on Kubernetes: Red Team Playbook
35mGenAI Application Security: Not Just Prompt Injection
22mAI in Kubernetes The Promise
18mCommit Secure AI: Embedding Security in Code, Pipelines, and Production
28mSecuring AI by design
28mAI for Everything After Code
25mSecuring AI Pipelines: Real-World Attacks on Kubernetes-Based AI Infrastructure