Securing AI Systems in Production

2026-05-16 · with PLACEHOLDER Guest, Head of ML Security

Prompt injection, data leakage, and agent abuse. What breaks when LLMs hit real users.

In this episode

• The LLM threat surface traditional appsec misses
• Defending RAG pipelines from data exfiltration
• Treating model tools as privileged APIs
• What 'AI red teaming' really involves

Transcript