Intermediate

RAG Systems & Vector Databases

Design production-ready Retrieval-Augmented Generation (RAG) pipelines. Master chunking strategies, vector index optimization, metadata filtering, and evaluation.

JK
Created by James Kariuki
4.7rating
950 learners enrolled
10 hours duration

What you'll learn

Implement smart document parsing and chunking strategies
Optimize vector indexing (HNSW, Flat) for query speed and recall
Build hybrid search engines (Semantic + Keyword search)
Set up query rewriting, expansion, and re-ranking
Evaluate RAG retrieval accuracy and response generation objectively

Course Content

Section 1: RAG Fundamentals
The RAG architecture: Retrieval, Augmentation, Generation
15 mins
Limitations of native LLM context windows
10 mins
Choosing your vector database: Pinecone, Pgvector, Qdrant
25 mins
Section 2: Ingestion and Processing Pipeline
Parsing complex documents: tables, headers, formatting
30 mins
Chunking strategies: recursive, sliding window, semantic
30 mins
Generating and storing vector embeddings
20 mins
Section 3: Advanced Retrieval
Query expansion and rewriting algorithms
25 mins
Re-ranking search results with Cohere / BGE Re-rankers
20 mins
Metadata filtering and hybrid search setup
35 mins

Your Instructor

JK

James Kariuki

Senior MLOps Engineer at AI Lab Nairobi

James designs data pipelines and scalable infrastructure for AI systems. He is an active contributor to LangChain and vector DB tools.

Prerequisites

  • Intermediate Python knowledge
  • Basic understanding of vector embeddings
  • Access to an LLM provider API key (OpenAI/Anthropic)
$85USD
Choose Payment Method
This course includes:
Full lifetime access
Access on mobile and desktop
Certificate of completion
Exercises & course resources