Mastering RAG Systems: A Hands-on Guide to Production-Ready AI

New
·

2 Weeks

·

Cohort-based Course

Master the skills to build and deploy production-ready RAG systems, combining AI and search for real-world, scalable applications.

Previously worked with

fast.ai
Stanford University
Anthropic
U.S. National Science Foundation
United Nations Foundation

Course overview

From RAG novice to building production-ready AI systems.

In this course, you'll gain the practical skills to design, build, and prototype Retrieval-Augmented Generation (RAG) systems that can scale for real-world applications. Starting with core concepts, you'll work through hands-on projects, mastering techniques to combine search and AI models for powerful, efficient systems. By the end, you'll be equipped to confidently deploy RAG systems in production environments, solving real business problems with cutting-edge AI.


What to expect in the course:


Week 1

Module 1: Introduction to Retrieval-Augmented Generation

-----------------------------------------------------------------------------------------

Welcome - Introduction to this course

~ Course Overview - What to Expect

~ Tools and techstack - eg. Python, LLM Providers, NBDev, and Qdrant

~ Prerequisites - Python Proficiency, Machine Learning Basics (JavaScript Helpful but Optional)

RAG Systems Overview - Setting the Stage

~ Defining RAG - The Core Idea Behind the System

~ Why RAG? - The Power of Combining Retrieval and Generation

~ Conceptual Flow of RAG Systems

Historical Evolution of RAG Systems

~ Development of Generative Models - A Brief History

~ Rise of Retrieval-Augmented Systems - How We Got Here

## QA-Retriever-Reader vs.

## QA-Retriever-Generator

RAG System Architecture - Key Concepts and Components

When to Apply RAG - Key Use Cases Explored

Challenges and Limitations of RAG Systems


Module 2: Architecture and Components of RAG Systems

-----------------------------------------------------------------------------------------

Understanding the RAG Pipeline - An Overview

Data Flow in RAG Systems - How Information Moves

The Retriever - The Heart of RAG Systems

~ Dense vs. Sparse Retrieval - Key Differences Explained (e.g., BM25, DPR, etc.)

~ Exploring Retrieval Methods for RAG Pipelines

~ Knowledge Stores - Understanding Vector Databases

The Generator Component - Creating Responses

~ Generation Models Overview - Powering the Generator

~ Input Representation - Preparing Data for Generation

Bringing It All Together - The Full RAG Workflow

Hands-On Practice - Setting Up a Basic RAG Pipeline


Module 3: Preparing and Ingesting Data for RAG Systems

-----------------------------------------------------------------------------------------

Parsing Raw Documents: The First Step to Understanding

Extracting Key Metadata from Documents

Document Chunking: Structuring Data for Retrieval

Embedding Document Chunks for Efficient Search (Transformers, OpenAI, Jina, Nomic, etc.)

Indexing Document Embeddings in a Vector Database


Week 2

Module 4: Building a Complete RAG Pipeline

-----------------------------------------------------------------------------------------

Implementing RAG Using Popular Frameworks

~ RAG with LlamaIndex

~ RAG with LangChain

~ RAG with Haystack

Building RAG from Scratch: A Comparison with Frameworks

Advanced RAG Techniques

~ Enhancing Retrieval with Query Expansion & Rewriting

~ Optimizing Results through Query Re-Routing

~ Improving Accuracy with Re-Ranking Strategies

~ Boosting Efficiency through Caching

~ Refining Retrieval Over Time by harnessing Feedback Loops

~ Exploring Other Advanced Retrieval Techniques (Raptor, Agentic RAG, Corrective RAG, HyDE etc.)

Integrating RAG: Front-End and Back-End Development

Hands-On Project: Build a Research Paper Chatbot from Scratch

🌶️ Bonus: Scaling RAG Systems



Module 5: Evaluation and Fine-tuning

-----------------------------------------------------------------------------------------

Evaluating RAG Systems: Metrics and Methods (Precision, recall, AP@k, quality, response time, MRR, NDCG, hit rate, human evaluation)

Optimizing LLMs for Enhanced Accuracy

When and How to Fine-Tune Your RAG Pipeline


Module 6: Observability and Cost

-----------------------------------------------------------------------------------------

Common Issues in LLM Applications (Low recall, noisy results)

Understanding Observability in LLM Applications (Logging, user feedback etc.)

LLM Monitoring vs. Observability: Key Differences (Latency, error rate, user engagement)


Practical Guide: Setting Up Observability for LLM Applications




Who is this course for

01

AI Engineers & Developers: For AI engineers/developers looking to master production-ready RAG systems combining search with AI models.

02

Data Scientists: Ideal for data scientists seeking to expand into AI by learning hands-on RAG techniques for real-world applications.

03

Tech Leads & Product Managers: Perfect for tech leads/product managers wanting to guide teams in building and deploying scalable RAG systems

What you’ll get out of this course

Introduction to RAG systems

Get a solid foundation in Retrieval-Augmented Generation (RAG), designed for novice learners. You'll grasp the basics of combining AI with search and build confidence in developing production-ready systems.

Build your first RAG system from scratch

Walk through step-by-step to create your first RAG system, even with no prior experience. This course covers all the fundamental concepts you need to start deploying real-world AI solutions.

Design production-ready RAG systems

Learn to design scalable Retrieval-Augmented Generation systems using search and AI, ready for real-world applications. Gain practical skills to manage system components and optimize performance.

Optimize RAG for real-world use cases

Learn how to fine-tune and optimize RAG systems for specific industries and use cases, from customer support to content generation. By the end, you'll be able to tailor AI solutions to meet diverse business needs.

This course includes

6 interactive live sessions

Lifetime access to course materials

In-depth lessons

Direct access to instructor

Projects to apply learnings

Guided feedback & reflection

Private community of peers

Course certificate upon completion

Maven Satisfaction Guarantee

This course is backed by Maven’s guarantee. You can receive a full refund within 14 days after the course ends, provided you meet the completion criteria in our refund policy.

Course syllabus

Week 1

Dec 7—Dec 8

    Oct

    26

    Live Session - Week 1 Session 1

    Sat 10/2610:00 PM—11:30 PM (UTC)

    Introduction to Retrieval-Augmented Generation

    6 items

    Oct

    27

    Live Session - Week 1 Session 2

    Sun 10/2710:00 PM—11:30 PM (UTC)

    Architecture and Components of RAG Systems

    7 items

    Preparing and Ingesting Data for RAG Systems

    5 items

Week 2

Dec 9—Dec 15

    Oct

    31

    [Optional] Office Hour

    Thu 10/3110:00 PM—11:00 PM (UTC)

    Nov

    2

    Live Session - Week 2 Session 1

    Sat 11/210:00 PM—11:30 PM (UTC)

    Building a Complete RAG Pipeline

    6 items

    Nov

    3

    Live Session - Week 2 Session 2

    Sun 11/311:00 PM—12:30 AM (UTC)

    Evaluation and Fine-tuning

    3 items

    Observability and Cost

    4 items

Week 3

Dec 16—Dec 20

    Nov

    7

    [Optional] Office Hour

    Thu 11/711:00 PM—12:00 AM (UTC)

What people are saying

        really liking the business vs technical perspective. As technicians we clearly see how LLM's can benefit not just businesses but the world, the difficulty for us in communicating that honestly and clearly. Your approach of evaluating performance, while looking at projects from different perspectives is helpful.
@AwakenwithoutCoffee

@AwakenwithoutCoffee

        Thanks for providing the production perspective in your videos. I believe this is unique to your channel, so keep doing that!
Egil Sandfeld

Egil Sandfeld

Engineer

Meet your instructor

Mehdi Allahyari

Mehdi Allahyari

Mehdi is a Chief AI Engineer, accomplished AI Researcher, and winner of Anthropic’s 2024 AI Developer Contest. With over a decade of experience in AI, LLM, NLP, and ML, he has led complex AI implementations across finance, technology, and gaming sectors. Previously a Computer Science professor at Georgia Southern University, Mehdi has earned over 2000 citations for his research and conference best paper awards. He has also collaborated with esteemed institutions like the National Science Foundation on advanced AI solutions. Mehdi is the published author of A Practical Approach to Retrieval-Augmented Generation Systems (Oct 2023) and co-host of the TwoSetAI YouTube channel, where he shares cutting-edge RAG techniques.


Follow him on Linkedin or YouTube

Angelina Yang

Angelina Yang

Angelina is a seasoned full-stack data scientist and AI startup founder. She is a two-time fast.ai fellow under Jeremy Howard and holds five patents in machine learning and natural language processing. As a former VP of data and ML, Angelina has built and led high-performing data science teams across industries like finance, fintech, and gaming. She is also a published author and Substack writer, with works on Retrieval-Augmented Generation and Machine Learning Interviews. Winner of Anthropic’s 2024 AI Developer Contest, Angelina continues to share her expertise as co-host of the TwoSetAI YouTube channel.


Follow her on Linkedin or YouTube

A pattern of wavy dots

Join an upcoming cohort

Mastering RAG Systems: A Hands-on Guide to Production-Ready AI

Cohort 1

$1,500

Dates

Dec 7—20, 2024

Payment Deadline

Oct 25, 2024

Course schedule

4 - 6 hours per week

  • Live Sessions on Saturdays & Sundays

    6:00pm - 7:30 pm EST

    Concept and code walk throughs of building a RAG system end-to-end.

  • Starting Oct 26, 2024

    For 2 weeks


  • Weekly Projects and Office Hours

    3 hours per week

    Office hours: 1 hour per week

    Weekly project: 2 - 4 hours per week

  • Offline Contents

    1 hour per week

    Good reads to take home for your busy schedule.

Free resource

Ebook: A Practical Approach to Retrieval Augmented Generation Systems

This free eBook provides a foundational understanding of Retrieval-Augmented Generation (RAG) systems. It covers real-world applications, such as using RAG for interacting with PDF documents, and explores various frameworks like LLamaIndex, LangChain, and Haystack. It also addresses practical challenges in deploying RAG systems. While this book lays the groundwork, our course will offer the latest updates and advanced techniques. Download it to build a solid base before diving deeper with us.

Get this free resource

Learning is better with cohorts

Learning is better with cohorts

Active hands-on learning

This course builds on live workshops and hands-on projects

Interactive and project-based

You’ll be interacting with other learners through breakout rooms and project teams

Learn with a cohort of peers

Join a community of like-minded people who want to learn and grow alongside you

Frequently Asked Questions

A pattern of wavy dots

Join an upcoming cohort

Mastering RAG Systems: A Hands-on Guide to Production-Ready AI

Cohort 1

$1,500

Dates

Dec 7—20, 2024

Payment Deadline

Oct 25, 2024

$1,500

2 Weeks