Orchestrators and subagents, the Model Context Protocol, RAG architecture, and the workflow patterns behind production agentic systems.
Multi-agent architecture patterns: orchestrator/worker, peer networks, hierarchical delegation, and the trust boundaries that emerge between agents.
Model Context Protocol: how it works, why it matters, building MCP servers, and why MCP security is the hottest attack surface in AI right now.
Retrieval-Augmented Generation from first principles: chunking, embedding, retrieval strategies, re-ranking, and the security implications of knowledge base poisoning.
Building production agentic systems: task planning, workflow routing, parallel execution, human-in-the-loop patterns, and failure recovery that actually works.