Speaking at DEF CON: Interview with Andra Lezza

In this insightful interview from DEF CON 33, host Hal Vetagoth sits down with application security and AI researcher Andra Lezza to discuss her work securing modern artificial intelligence systems. Opening at 00:00, Andra introduces her talk on securing autonomous AI agents and multi-agent workflows (MTP security), outlining critical threat modeling principles, supply-chain protections, and architectural guardrails required when deploying autonomous agents in production environments.

At 01:02, Andra explores the steep learning curve in AI security, candidly reflecting on the balance between industry hype and genuine security concerns. She contrasts the enormous scale of DEF CON in Las Vegas—with massive crowds spanning casino convention centers—against European cybersecurity conferences.

At 02:00, she previews an encore presentation at the AppSec Village focusing on securing complex multi-agent architectures. Closing at 03:02, Andra reflects warmly on her first DEF CON experience connecting with the OWASP community, praising the friendly, supportive, and collaborative ethos that defines the convention. (AI summary — Gemini Flash)