Udemy
    •  
    •  
    •  
    •  
    •  
    •  
    •  
    •  
Subscribe
Turn what you know into an opportunity and reach millions around the world.
Learn More
Your cart is empty.
Keep shopping
RAG 101: Build AI Systems with Your Own Data
Rating: 5.0 out of 5(3 ratings)
11 students

RAG 101: Build AI Systems with Your Own Data

Learn RAG from scratch: chunking, embeddings, vector databases, retrieval, prompting, and advanced RAG architectures.
Created bySafaa Khadim
Last updated 4/2026
English

What you'll learn

  • Learn RAG from scratch, step by step, with the intuition behind every building block and design choice.
  • Understand embeddings, chunking, indexing, and retrieval in depth—what they do, why they work, and how to tune them.
  • Build a complete end-to-end RAG system and explore multiple architectures and techniques for better accuracy.
  • Compare key RAG techniques (semantic search, hybrid search, reranking, filtering, MMR) and know when to use each one.
  • Improve RAG quality by diagnosing retrieval failures, hallucinations, and weak grounding using practical methods.

Course content

8 sections • 25 lectures • 2h 41m total length

  • Introduction7:44

Requirements

  • Basic Python knowledge is recommended for the implementation and hands-on sections.
  • No prior knowledge in Artificial Intelligence or Machine Learning is required.
  • All RAG concepts are explained from scratch, step by step.
  • The focus is on intuition and understanding, not advanced math or theory.

Description

This course contains the use of artificial intelligence.

Large Language Models are powerful, but on their own, they lack access to your private data and cannot be trusted to answer reliably. This is where Retrieval-Augmented Generation (RAG) becomes essential.

In RAG 101: Build AI Systems with Your Own Data, you will learn how to design and build reliable, end-to-end RAG systems from scratch, with a strong focus on intuition, architecture, and real-world best practices.

This course starts from first principles and gradually takes you to advanced RAG architectures. Every concept is explained clearly: not just how things work, but why they work.

You will begin with chunking fundamentals, learning how to split documents effectively and how chunk size and overlap directly impact retrieval quality. You will then dive into embeddings, understanding how text is transformed into vectors and how semantic similarity actually works.

Next, you will explore vector databases, including indexing strategies, metadata usage, and how modern vector stores power fast and accurate semantic search. From there, you will master the retrieval phase, covering similarity metrics, static and dynamic Top-K strategies, filtering, and advanced techniques such as MMR, BM25, and reranking.

You will also learn prompt construction for RAG, including how to combine user queries, retrieved context, and instructions using advanced prompting techniques like few-shot prompting, query rewriting, and guardrails.

Finally, the course introduces advanced RAG architectures, including Graph RAG, Multi-Vector RAG, and Agentic RAG, helping you understand when and why to move beyond vanilla RAG.

This course is designed for developers, engineers, and technical professionals who want to move from basic RAG prototypes to production-ready AI systems, with confidence and clarity.

Who this course is for:

  • Developers and engineers who want to build LLM applications using their own data
  • Data scientists and AI practitioners looking to understand RAG from first principles
  • Technical professionals, consultants, and product teams working with AI solutions
  • Beginners in AI who want a clear, intuitive introduction to RAG
  • Anyone curious about how to make LLMs reliable, explainable, and grounded in real data

Report abuse