From Raw Documents to Graph Intelligence: Building AI Pipelines with LangChain, Neo4j, and GitHub

Standard vector-search tutorials often oversimplify AI development, leaving developers stuck when trying to process messy, real-world documents. In this hands-on, 2-hour workshop, we will bridge the gap between "dummy" chatbot examples and true production-grade backend engineering.
You will learn how to build an end-to-end intelligent document processor. First, we will construct a multimodal pipeline using Python and LangChain to parse, classify, and extract structured data from unstructured formats (such as complex PDFs, forms, and tables). Next, we will step beyond flat vector storage and model our extracted metadata, entities, and document relationships into a highly queryable Knowledge Graph using Neo4j.
To maximize our 2-hour window and bypass environmental hurdles, we will avoid the common trap of overplanning and keep our coding scope razor-sharp. Using GitHub as our workspace, every participant will fork and clone a production-ready starter repository containing a comprehensive README and a step-by-step implementation guide, which we will extend live during the session.
By the end of this workshop, you’ll walk away with a deep understanding of hybrid AI-graph architectures, real-world parsing strategies, and a fully functional portfolio project pinned directly to your GitHub profile. Perfect for intermediate Python developers and aspiring AI engineers who want to build connection-aware backend systems.