AI and Agentic security - build, break and secure | Ep. 90 скачать в хорошем качестве

AI and Agentic security - build, break and secure | Ep. 90

AI is already in production everywhere, but most teams are still figuring out how to secure it. Prompt injection, poisoned models, leaking data, misbehaving agents… these are problems people are dealing with today, not in research papers. In this episode, we’ll talk honestly about what it actually takes to secure AI systems in the real world. We’ll cover: how AI/Agents are built, what AI attacks look like in practice, and what actually stops them, how to think about AI red teaming and guardrails without slowing teams down, securing agentic (MCP / A2A / Skills) as agents become more autonomous and where AI security fits in a modern DevOps pipeline. Bonus content! We'll also introduce the AI Security Learning Journey—a hands‑on path built for developers and security engineers. If you’re building with AI, operating it, or responsible for securing it, this session is meant for you