The AI Security Engineer Lead owns the engineering of security controls for agentic AI: the agent harnesses that house and orchestrate agents, tools, the MCP servers and APIs that agents call, agent-to-agent (A2A) interactions, and the AI and API gateways that sit in the path. Define guardrails against prompt injection (direct and indirect), tool misuse, excessive agency, data leakage, and model or supply-chain compromise, aligned to OWASP Top 10 for LLM Applications, OWASP Agentic AI guidance, and MITRE ATLAS.