10/26/2026
This 45-minute guest lecture introduces AI agents and multi-agent systems from an engineer’s perspective, walking through the nine layers of the agentic AI stack: LLM, context, memory, RAG and vector databases, tools, orchestration, observability and evaluation, and security and governance. It also covers practical ways to control cost, including prompt caching, context trimming and matching the model tier to the work. The session ends with a live demo of a four-agent pipeline showing which steps run in sequence, which run in parallel, and why clear agent boundaries matter.