Building intelligent, secure, and scalable autonomous systems
summary
Runs LLMs in a loop as enterprise agents that perceive code or docs, reason a plan, then execute tools; demonstrates a real Java 8→21 upgrade agent that builds, runs Gradle, and requires human plan approval; enforces sandboxed containers and VMs plus file system bubblewrap, least‑privilege JSON queries instead of raw SQL, comprehensive logging and audit trails, continuous red‑teaming, MCP via remote MCP servers, multitenant control plane and per‑agent cost/observability.
AI-written summary · about
chapters (9)
- 0:00Enterprise agents introduction
- 3:08Coding agent demo
- 5:14Security for agent systems
- 8:27Audit trails and red teaming
- 10:35Human oversight and monitoring
- 13:43LLMs to enterprise data
- 16:51MCP server integration demo
- 17:54Scaling agents with control plane
- 19:43Summary
related talks
11mBuilding Secure Backend Services with AI Agents
40mAgents in Action: Advancing Open Source AI Missions
34mAI Agents: Transforming Autonomous Decision-Making
50mIntelligent Agents, the New Frontier of LLMs
46mDemystifying AI Agents
54mGenAI Grows Up: Building Production-Ready Agents on the JVM
28mAgentic Development at Enterprise Scale: RAG, MCP & Guardrails
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…