How Modern AI Systems Really Find Answers: Build GraphRAG Applications

This hands-on workshop takes you beyond basic RAG to build an explainable financial advisor using Neo4j, knowledge graphs, Cypher, public filings, news, and LLM agents. Led by bestselling author and Chief Scientist at GraphAware, Dr. Alessandro Negro, you'll build the knowledge graph up progressively, module by module, until it becomes the single source of truth that grounds an agentic RAG system — one that combines structured and unstructured data to answer questions vanilla vector search misses: multi-hop reasoning, entity relationships, change over time, and source-traceable answers. Everything runs on real, complex datasets and a reliable graph database (Neo4j), and the code shared with you is written to be applicable in production, not just in a demo.

REGISTER HERE: https://luma.com/z3mtzgrf?coupon=STANDARD20&utm_source=organic

What this workshop is about

• Progressively build a knowledge graph that becomes your single source of truth
• Ground an agentic RAG system that combines structured and unstructured data
• Move beyond basic vector RAG into graph-backed, agentic reasoning
• Ingest corporate filings, news, and public financial data
• Extract entities and relationships in multiple verified steps — not one risky single shot (the approach behind Microsoft GraphRAG)
  Use agentic retrieval, graph navigation, and text-to-Cypher
• Work with real, complex datasets and hit the real ceilings production systems face
• Answer complex financial questions with traceable, explainable reasoning

Who this is for

• AI engineers building RAG and LLM applications
• ML engineers moving into graph-backed AI systems
• Data scientists working with complex document-heavy data
• Software architects designing explainable AI applications
• LLM app developers building production-grade assistants
• Technical consultants working on financial, legal, compliance, or enterprise AI use cases

Tools and frameworks you will learn

• Neo4j as a reliable graph database for knowledge graph storage and querying
• Multi-step, verified LLM entity and relationship extraction (agentic, à la Microsoft GraphRAG)
• Cypher for graph queries
• Docling for document ingestion
• LLM agents for iterative retrieval and reasoning
• Vector search for semantic retrieval
• Keyword and Lucene-style search for hybrid retrieval
• Text-to-Cypher for natural language graph querying
• LLM-based entity and relationship extraction
• Entity resolution and graph enrichment techniques
• LLM-as-judge and evaluation workflows
  

Why attend this NOW

• Build a production-ready, scalable GraphRAG architecture — not toy demos
• See how a knowledge graph acts as a single source of truth for agentic RAG
• Learn multi-step, verified entity/relationship extraction instead of fragile single-shot extraction
• Work with real, complex datasets and see where real production limits appear (e.g. API/download rate ceilings) — and how the shared code is built to handle them
• Learn directly LIVE from bestselling author and data scientist
  - Get certified and receive complete recording and resources
• Build a complete project, not isolated demos
• See how to combine vector search, knowledge graphs, Cypher, and agents
• Understand design decisions behind each architecture layer
• Get explainability, traceability, and evaluation built into the workflow
• Go beyond free tutorials that only cover basic RAG or toy examples
• Learn a reusable architecture transferable to finance, legal, compliance, and healthcare
• Prepare for the next wave of enterprise AI systems: graph-backed, agentic, and auditable

What you will get

• Certificate of completion
• Full HD recording
• Presentation PPT
• Ready-to-use GraphRAG architecture blueprint
• Full working codebase shared upfront
• Neo4j knowledge graph schema examples
• Agentic retrieval workflow patterns
• Text-to-Cypher prompt and guardrail patterns
• Multi-step entity & relationship extraction pipeline with verification and constraints
• Evaluation checklist for GraphRAG systems
• Production-readiness checklist for financial AI advisors
• Reusable design framework for graph-backed LLM applications

Pre-requisites

• Comfortable with Python
• Basic experience with LLMs
• Familiarity with RAG concepts
• Some understanding of embeddings and vector search
• No prior Neo4j or Cypher experience required
• No finance domain expertise required

About the speaker

Dr. Alessandro Negro is Chief Scientist at GraphAware, where he leads science and technology work around Hume, a mission-critical knowledge graph analytics platform. He is a bestselling author of celebrated graph-powered machine learning, knowledge graphs, and LLM books. His work focuses on knowledge graphs, LLMs, NLP, graph-aided search, recommendation systems, and explainable AI, with deep experience applying graph technologies to real-world enterprise and investigative use cases.