Skip to main content
This course takes you from your first LLM-powered agent to robust, production-ready systems. Across seven modules you’ll build an agent, instrument it for observability, design architectures with tools and MCP, ground it with agentic RAG, evaluate its behavior, and monitor it in production. Audience: engineers building or operating agents. Prerequisites: familiarity with LLM evaluation is helpful but not required.

​
Modules

Module 1: Introduction to agents

What agents are, how they work, and the core building blocks you’ll use throughout the course.

Module 2: Agent engineering and observability

Build an agent and instrument it with traces and spans to see what it does at every step.

Module 3: Agent architectures and frameworks

Common agent architectures and how the major frameworks compare.

Module 4: Tools and MCP

Give agents tools and connect them to external systems with the Model Context Protocol.

Module 5: RAG and agentic RAG

Ground agents in your data with retrieval, and let them drive retrieval agentically.

Module 6: Agent evaluation

Measure agent quality with code-based and LLM-as-a-judge evaluations.

Module 7: Post-deployment and monitoring

Monitor agents in production and keep improving them with real-world feedback.