Generative AI and LLMs Section 116
Agentic Frameworks
The frameworks people build real agents with — LangGraph, MCP, multi-agent systems — and how to evaluate what you built.
12 of 12 lessons published Three reading levels on every lesson
Lessons in order
Work top to bottom. Each lesson assumes the one above it.
- LangGraph — agents as graphs
- State and checkpoints in LangGraph
- CrewAI — role-based agent teams
- AutoGen — conversational multi-agent systems
- MCP — the Model Context Protocol
- DSPy — programming, not prompting
- smolagents — code-acting agents
- Designing good tools for agents
- Agent memory
- Human-in-the-loop agents
- Evaluating agents
- Orchestrating multiple agents