Course Wiki

LangChain

The framework module: what an LLM abstraction layer gives you and what it hides.

Module 6 introduces LangChain as the widely used abstraction layer over LLM applications: chains compose prompts, models, and retrievers; document loaders and text splitters feed the knowledge base; vector stores and retrievers are pluggable backends. The course shows the same RAG pipeline from earlier modules re-expressed in LangChain’s vocabulary.

The teaching stance is balanced: abstractions speed up standard patterns and make component swaps easy, but the course insists you first understand what the chain does under the hood — because when output quality drops, debugging happens at the pipeline level the framework wraps.

Taught in

• LLM Zoomcamp

  • Module 1: Agentic RAG
    • The Agentic Loop
    • ToyAIKit
    • Other Frameworks
  • Module 6: Best Practices
    • Hybrid Search with LangChain
    • Next Steps ## Related concepts

• RAG

• Vector Search

• Hybrid Search

• Reranking

• LLM Evaluation