Description

This RAG training course teaches you how to build Retrieval-Augmented Generation systems that allow AI to pull from your own documents and data sources for accurate, grounded answers. You will learn vector databases, embedding strategies, chunking techniques, hybrid search, evaluation methods, and how to deploy production-grade RAG systems for customer support, internal knowledge bases, and AI-powered search experiences.

Course Content

Module 1: Why RAG?

  • The limitations of LLMs on their own
  • When RAG beats fine-tuning
  • Real-world use cases for RAG

Module 2: RAG Architecture Fundamentals

  • The components of a RAG system
  • Retriever, ranker, and generator roles
  • The RAG pipeline end-to-end
  • Common architecture patterns

Module 3: Embeddings and Vector Search

  • What are embeddings and how they work
  • Choosing embedding models
  • Vector similarity metrics
  • Tradeoffs between accuracy and speed

Module 4: Vector Databases

  • Pinecone, Weaviate, Chroma, pgVector
  • Self-hosted vs. managed options
  • Indexing strategies and scalability
  • Hybrid search: vector %2+ keyword

Module 5: Chunking and Data Preparation

  • Chunking strategies for different content
  • Preprocessing PDFs, tables, and images
  • Metadata and filtering
  • Handling updates and deletions

Module 6: Advanced RAG Techniques

  • Query rewriting and expansion
  • Re-ranking with cross-encoders
  • Multi-query and ensemble retrieval
  • Agentic RAG patterns

Module 7: Evaluation and Quality

  • Metrics: precision, recall, faithfulness
  • Building evaluation datasets
  • Detecting and reducing hallucinations
  • Tools: Ragas, LangSmith

Module 8: Production RAG

  • Caching and performance optimization
  • Cost management at scale
  • Citations, grounding, and trust
  • Build a RAG system for your own data

Duration: 5 – 7 weeks

Hi, How Can We Help You?