Agent SDKs & Frameworks · TRACK 03/05
LangChain & LangGraph
The biggest ecosystem in LLM tooling, demystified.
// THE TRACK
01 · START HERELangChainCut through both the hype and the backlash: understand what LangChain actually provides and the kinds of projects where it genuinely helps.BEGINNERLangGraphUnderstand how LangGraph models an agent as a graph of nodes and shared state — and why that explicitness matters for loops, branching, checkpointing, and human oversight.INTERMEDIATELangChain vs LangGraphGet a clear rule of thumb for when a simple chain is enough and when your workflow has earned a stateful graph — plus how the two work together.INTERMEDIATELangSmithLearn to read a LangSmith trace, find the step where your chain went wrong, and set up your first evaluation dataset.INTERMEDIATELCELUnderstand LangChain Expression Language — how the pipe operator composes runnables and gives you streaming, batching, and async out of the box — so LangChain code finally reads cleanly.INTERMEDIATEPersistence & CheckpointersLearn how LangGraph's checkpointers persist graph state — enabling pause/resume, human-in-the-loop approvals, time-travel debugging, and memory that survives across sessions.INTERMEDIATEMigrating Off LangChainGet a practical, incremental plan for moving off LangChain — extracting your prompts and tool logic, replacing one layer at a time — without a risky big-bang rewrite.INTERMEDIATE