LangChain
LangChain and LangGraph — the orchestration layer for serious agent workflows.
Once you've outgrown direct API calls and need to build agents that reason across multiple steps, use tools, maintain state, and recover from errors — you need an orchestration framework. LangChain is the most widely adopted one, and LangGraph (its graph-based extension) is what powers the most sophisticated AI agents being built today.
In this track, I explain why you need a framework in the first place (it's not just hype — it solves real problems), how LangChain's LCEL (LangChain Expression Language) chains components together, and how LangGraph lets you build cyclic, stateful agents that can retry, branch, and loop. Multi-agent swarms — where specialized agents collaborate — are also covered.
If you've ever had an agent that worked in the notebook but broke in production because it lost track of what step it was on, LangGraph's state persistence and conditional edges are the solution. This track gives you the patterns to build agents that are actually reliable.
📚 Learning Path
- Why orchestration frameworks exist
- LangChain LCEL chains and runnables
- LangGraph: stateful cyclic agents
- Multi-agent swarm coordination
- Production patterns: retry, fallback, checkpointing