The 5 Levels of RAG: From Naive Retrieval to Agentic RAG
Bashiri Smith · Facebook reel · 2026-10-12 · 0:53 · 10,323 views · Open on Facebook
Topics: Retrieval-Augmented Generation (RAG), Embeddings & Vector Databases, AI Agents, Tool Use & MCP · Level: intermediate
Summary
Bashiri Smith describes five levels of Retrieval-Augmented Generation, counting down from level 5 to level 1. They go from naive embed-and-retrieve pipelines to smarter chunking, hybrid search with reranking, advanced RAG with query rewriting and citation checks, and finally agentic RAG. For a first pipeline he points to ChromaDB and the LangChain Quickstart, and for chunking he points to LangChain's text splitters.
Key points
- Level 5 – Naive RAG: embed your documents into a vector store, retrieve the top-k chunks and put them into the prompt. ChromaDB plus the LangChain Quickstart gets a first pipeline working.
- Level 4 – Smarter chunking: replace blind fixed-size splits with recursive splitting, section-aware chunks and metadata so retrieval finds the right context. LangChain's text splitters cover this.
- Level 3 – Reranking and hybrid search: combine dense vector search with BM25 keyword search and merge the results with Reciprocal Rank Fusion (RRF).
- Level 2 – Advanced RAG: query rewriting, multi-hop retrieval and citation verification, so every claim is grounded in a source and checked.
- Level 1 – Agentic RAG: agents decide what to retrieve and when, retry when results are poor, and correct themselves at scale.
- The creator's free RAG Guide is in a Google Drive folder linked in the caption.
Resources mentioned
- BASWE.AI: RAG Guide · pdf · drive.google.com · free
The creator's RAG guide, shared as a Google Drive folder. - BASWE.Ai Engineer (Skool community) · community · skool.com · paid
The creator's paid community and program, with an AI learning roadmap (including the full ops and evaluation track), daily calls with engineers and recruiters, resume and portfolio help, and a job-search pipeline.
Also in: Basic vs Production AI Engineering: RAG, Agents, and Context over Prompts (Bashiri Smith on Facebook · notes), ChatGPT Prompt to Generate Standout Portfolio Project Ideas (Bashiri Smith on Facebook · notes), Vector Database vs. Graph Database: When to Use Each (and Both for RAG) (Bashiri Smith on Facebook · notes), Get More Interviews: Prove You Built Your Projects With Video Walkthroughs (Bashiri Smith on Facebook · notes) and 93 more - Chroma · tool · trychroma.com · free
The easiest open-source vector database to start with. - LangChain Quickstart · docs · docs.langchain.com · free
LangChain's getting-started guide for building LLM applications, including RAG pipelines. - LangChain · tool · github.com · free · recommended by both Bashiri Smith & Melvin Vivas
Named in the prompt as a beginner-friendly but powerful AI framework for building AI-powered projects.
Also in: Find Hidden AI Engineering Jobs by Searching for Skills, Not Job Titles (Bashiri Smith on Facebook · notes), Graph Engineering Explained: Nodes, Edges, and State for AI Agent Workflows (Bashiri Smith on Facebook · notes), SWE-to-AI Engineer Plan for 2027: LLMs, RAG, Agents, Evals, Job Search (Bashiri Smith on Facebook · notes), Basic RAG Pipeline in 60 Seconds: From Documents to Grounded Answers (Bashiri Smith on Facebook · notes) and 6 more
Try this
- Build a naive RAG pipeline with ChromaDB and the LangChain Quickstart.
- Upgrade the chunking with recursive splitting, section-aware chunks and metadata.
- Add hybrid search (dense vectors plus BM25) and merge the results with Reciprocal Rank Fusion.
- Comment "RAG" on the post to get the creator's full RAG roadmap.
- Build one RAG system in stages, from naive retrieval to smarter chunking, hybrid search with RRF, query rewriting with citation checks, and finally an agentic retriever.
More in Retrieval-Augmented Generation (RAG)
- LlamaIndex OCRia-1: A New Vision-Language Model for Parsing Document Tables
- Basic vs Production AI Engineering: RAG, Agents, and Context over Prompts
- Production RAG: Check Retrieval, Answer Support, and Unseen Questions
- Entity Linking: Connecting Text Mentions to Structured Records
- 5 RAG Metrics: Top-k, Recall@k, Precision@k, Top-p and Cost per Request