Skip to Content
View all events

GraphRAG Bootcamp

Published by O'Reilly Media, Inc.

Intermediate content levelIntermediate

Building grounded and reliable AI applications with LLMs and knowledge graphs

This live event utilizes Jupyter Notebook technology

Course outcomes:

  • Understand when and how knowledge graphs improve LLM-based applications compared to traditional RAG approaches.
  • Design and develop GraphRAG systems using different architectural patterns tailored to specific use cases.
  • Design and build knowledge graphs that effectively support GraphRAG applications.
  • Evaluate, diagnose, and systematically improve the behavior of GraphRAG systems.

Course description

Join expert Panos Alexopoulos for a 2-day immersive bootcamp on designing and building GraphRAG systems, namely hybrid AI applications that combine large language models with knowledge graphs to deliver more grounded, reliable, and explainable results. Through real-world use cases, practical examples, and hands-on exercises, you’ll learn how to design GraphRAG architectures tailored to your needs, as well as how to build knowledge graphs that effectively support them. In addition, you will learn how to evaluate the behavior of GraphRAG systems in practice, identify common failure modes and their root causes, and systematically improve their performance.

NOTE: With today’s registration, you’ll be signed up for both sessions. Although you can attend either of the sessions individually, we recommend participating in both.

What you’ll learn and how you can apply it

  • Decide when a GraphRAG approach is more suitable than traditional RAG for your AI application
  • Design GraphRAG architectures that align with your use case requirements and constraints
  • Build knowledge graphs that effectively support retrieval, context construction, and reasoning
  • Integrate graph-based retrieval into LLM workflows to improve grounding and consistency
  • Evaluate the behavior of GraphRAG systems and identify their weaknesses
  • Diagnose common failure modes and trace them back to issues in data, retrieval, or generation
  • Apply systematic improvements to make your GraphRAG systems more reliable and robust

This live event is for you because...

  • You’re an aspiring or practicing AI engineer, data scientist, or developer working with LLM-based applications
  • You’re building or experimenting with RAG systems and want to make them more effective and reliable.

Prerequisites

  • Basic familiarity with large language models and LLM-based applications
  • Some experience with Python and Jupyter Notebooks (ability to run and modify simple code examples)
  • No prior experience with knowledge graphs is required

Recommended preparation:

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

Day 1

From LLMs to GraphRAG (60 minutes)

  • Core limitations of LLMs
  • How RAG addresses these limitations
  • How traditional RAG works and what are its limitations
  • What are knowledge graphs how they improve traditional RAG
  • Q&A
  • Break

Working with Neo4j (60 minutes)

  • Representing, storing and querying knowledge graphs in Neo4j according to the property graph model;
  • Neo4j GraphRAG libraries
  • Jupyter notebook:
  • Create a Neo4j sandbox and showcase basic data creation, manipulation, and access queries
  • Create and run a very simple GraphRAG pipeline
  • Q&A
  • Break

Developing GraphRAG systems (60 minutes)

  • Common GraphRAG architectural patterns and when to use each
  • Jupyter notebook:
  • Implement a GraphRAG career advice system
  • Q&A
  • Break

Day 2

Day 1 recap (15 minutes)

  • Presentation: Recap of day 1 highlights and key points
  • Q&A

Evaluating GraphRAG systems (60 minutes)

  • Choosing evaluation dimensions and metrics
  • Creating evaluation datasets
  • Computing and interpreting evaluation results
  • Diagnosing failure modes and determining improvement actions.
  • Jupyter notebook:
  • Evaluate the career advice GraphRAG system using DeepEval and RAGAS
  • Q&A
  • Break

Building knowledge graphs for GraphRAG (105 minutes)

  • Designing knowledge graph schemas
  • Populating knowledge graphs from unstructured data
  • Detecting and correcting quality problems in knowledge graphs
  • Jupyter notebook:
  • Build a career advice knowledge graph from job vacancies
  • Detect and correct quality problems in a knowledge graph using heuristics and AI/ML techniques.
  • Q&A

Your Instructor

  • Panos Alexopoulos

    Panos Alexopoulos works as head of ontology at Textkernel, in Amsterdam, Netherlands, where he leads a team of data professionals developing and delivering a large cross-lingual knowledge graph in the HR and recruitment domain. He’s the author of Semantic Modeling for Data: Avoiding Pitfalls and Breaking Dilemmas (O'Reilly) and has published more than 60 papers at international conferences and in journals and books. He has worked at the intersection of data, semantics, and software since 2006, contributing to building intelligent systems that deliver value to business and society. Born and raised in Athens, Greece, Panos holds a PhD in knowledge engineering and management from National Technical University of Athens and is a regular speaker and trainer in both academic and industry venues.

    linkedinXlinksearch

Skills covered

  • Large Language Models (LLMs)
  • Unified Modeling Language (UML)