RAG Architecture Simulation

Identify your external data sources (PDFs, Wikis, Databases) that the LLM is not trained on.
What is RAG? (Video) 📺

Extract text from your sources. Use loaders (like LangChain document loaders) to read the raw text.

Split long text into smaller, meaningful chunks to fit into the context window.
Chunking Strategies 📺

Convert text chunks into vector representations (arrays of numbers) using an embedding model (e.g., OpenAI, HuggingFace).

Save your vectors in a DB like Pinecone, Chroma, or Weaviate for efficient similarity search.

Accept the question from the user in natural language.

Embed the user's query and find the most similar text chunks ("Top K") from your vector database.
Vector Search Explained 📺

Combine the user's query with the retrieved text chunks into a single prompt for the LLM.

Pass the augmented prompt to the LLM to generate the final answer based on the retrieved facts.

Review the answer for accuracy and hallucinations. Iterate on chunking or retrieval if needed.