modiashu.dev

RAG Pipeline with LangChain and OpenAI

Python · LangChain · OpenAI · Vector Search · Jupyter

View source →

Overview

A hands-on RAG implementation covering the full retrieval-augmented generation loop:

  • Chunking — splitting source documents with overlap strategies
  • Embedding — vectorizing chunks for semantic search
  • Retrieval — similarity search against the vector store
  • Generation — grounding LLM responses in retrieved context

Related writing: my RAG evaluation piece (“Text Summarization Evaluation Paradox”) explores how evaluation methods hit the same token-limit constraints as summarization itself.

View source →