# Neo4j Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/neo4j
> Markdown URL: https://aitinkerers.org/technologies/neo4j.md
> Technology record last updated: 2026-02-25T17:31:55Z
> Generated: 2026-08-24T10:50:18Z

Neo4j is the world's leading native graph database, purpose-built for high-performance management and traversal of connected data.

Neo4j is the leading native graph database, leveraging the Property Graph Model (nodes, relationships, properties) for data storage and retrieval . It is an ACID-compliant, high-performance platform designed for managing highly connected data at scale (billions of nodes) . Queries execute using the declarative Cypher query language, which simplifies complex traversals that would cripple relational systems with numerous JOINs . This architecture delivers orders of magnitude performance improvements (often minutes to milliseconds) for relationship-based queries . Major use cases include fraud detection, recommendation engines, and knowledge graphs for AI, trusted by 84 of the Fortune 100 .

- Official technology site: https://neo4j.com
- Public AI Tinkerers demos and talks: 29
- Result page: 1 of 2

## Recent Public Talks and Demos

### [AI-Agents that learn from humans-in-the-loop](https://nurnberg.aitinkerers.org/talks/rsvp_YaF-2Es_QW0)

Enterprise AI Platform that not only integrates existing data sources and operational systems, but that enables AI Agents who can proactively identify the right employees to acquire and preserve missing knowledge or hand-off and delegate tasks.

- Event context: AI Tinkerers Nürnberg: April Meetup — 2026-04-22 — Nürnberg
- Public talk page: https://nurnberg.aitinkerers.org/talks/rsvp_YaF-2Es_QW0

### [Graph-Enhanced XGBoost: Beating Fraud Detection Baselines with Neo4j](https://saopaulo.aitinkerers.org/talks/rsvp_jIeBaJ1l7t8)

A graph-based financial fraud detection system built on Neo4j, using the IEEE-CIS dataset to model transactions, cards, devices, emails, and addresses as a connected network — and combining graph features with XGBoost to significantly outperform traditional tabular models.

- Event context: AI Tinkerers SP e Banco BMG — 2026-03-26 — São Paulo
- Public talk page: https://saopaulo.aitinkerers.org/talks/rsvp_jIeBaJ1l7t8

### [Miró: Synthetic Audience Analysis using LLM Agents and Graph Database](https://manizales.aitinkerers.org/talks/rsvp_QJ7Kot1_hP8)

In this demo, I will do a code deep-dive into "Miró", an engine I built to forecast the social and critical reception of upcoming books using synthetic readers. Instead of a product pitch, I will focus entirely on the technical architecture and the integration layer between LLMs and Graph Database. I'll walk through the code live, showing: Agent Generation Pipeline: How I parse static PDFs containing psychological profiles and translate them into "Synthetic Reader" nodes with their respective master prompts. The Predictive Engine: The orchestration code that drives how agents "read" the book, interact with each other, and how these interactions continuously update the graph database state. Graph-based Memory Management: A look into how I solved the challenge of persistent agent memory by dynamically creating complex relationships (such as CHATTED_ABOUT) and appending conversation histories as extendable edge properties using RAG. Analysis Dashboard: A quick look at how the Python backend consumes this dynamic graph network to feed an interactive react frontend, rendering resonance, friction, and abandonment connections.

- Event context: 🚀 ¡14vo Encuentro de AI Tinkerers Manizales! 🤖 — 2026-03-25 — Manizales
- Public talk page: https://manizales.aitinkerers.org/talks/rsvp_QJ7Kot1_hP8

### [Multi-Agent Fraud Detection: When LLMs Argue About Bitcoin Laundering](https://toronto.aitinkerers.org/talks/rsvp_l2U61QefZDs)

I built a real-time Bitcoin fraud detection system where three LLM agents debate whether a transaction is fraudulent: - Agent 1 (Prosecutor): Uses Graph RAG (Neo4j) to find suspicious network patterns and builds a case for fraud - Agent 2 (Defense): Searches for legitimate explanations and challenges the prosecutor's claims - Agent 3 (Judge): Reviews both arguments and makes the final verdict The Live Demo: 1. A suspicious transaction streams in via Kafka 2. Prosecutor Agent queries the Neo4j graph and constructs a fraud case using Gemini 3. Defense Agent counter-argues with alternative explanations 4. Judge Agent renders a verdict with confidence scoring 5. I pop the hood and show: the exact prompts, the Cypher graph queries, and the decision logic Technical Deep Dive: - How I structure multi-hop graph context for LLM reasoning - Prompt engineering to prevent agent "hallucination" on graph data - Latency battles: why I moved from Gemini Pro to Flash and added prompt caching - The surprising failure modes: when agents agree too quickly vs. when they hallucinate connections

- Event context: AI Tinkerers Toronto - January 2026 Meetup at Google! — 2026-01-29 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_l2U61QefZDs

### [Train Guard](https://paris.aitinkerers.org/talks/rsvp_JKBhzY72aQo)

Taking information from accident reports of train transportation (PDF) and preprocessing the data to create graph database to understand the relationships between causes and consequences.

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_JKBhzY72aQo

### [Neo4j Live: Inside StrangerGraphs – Predicting Season 5 with Graph Intelligence](https://paris.aitinkerers.org/talks/rsvp_3nhfxvOKTRg)

Join us as we break down StrangerGraphs, a prediction graph built from Reddit fan theories, Neo4j AuraDB, GPT-5 analysis, and GraphRAG-powered agents to explore what the Stranger Things community got right in past seasons – and what they might reveal about Season 5. Key Highlights: – Reddit prediction mining + GPT-5 accuracy scoring – Leiden clustering to find high-signal predictor communities – Season 5 predictions extracted from accuracy-based hubs – AuraDB + GraphRAG powering character-aware AI agents

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_3nhfxvOKTRg

### [Have questions about your building? Why not ask it directly in your language!](https://paris.aitinkerers.org/talks/rsvp_MZOL3K2NedI)

1. Core Problem - France PLU/PLUi Urban Regulation = huge, fragmented, constantly updated PDFs + annexes. - Hard to search, slow to analyze, high risk of late non-compliance. 2. Data Pipeline - Automated PDF - Normalizes zoning rules into a clean structure. - Everything stored in a Neo4j graph (zones, parcels, articles, files, images). 3. Knowledge Graph - Use LLM to generate Cypher-based retrieval used for Neo4j. - Full-text + vector search for hybrid rule lookup. - Each rule answer linked to exact source documents. 4. Multi-Agent System (OpenAI + LangGraph) Agents: - Regulation Agent (rules + citations) - Surveyor Agent (parcel + context) - Model Agent (auto model generation) - BIM Agent (FAR, height, envelope) - Supervisor (orchestration + verification) Workflow triggered by one address + one question. 5. 3D & Calculations - Automatic parcel + building context reconstruction. Outputs: 3D model, rendered images, feasibility metrics. 6. User Interaction - Chat interface powered by GPT.

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_MZOL3K2NedI

### [NeuroGraph](https://paris.aitinkerers.org/talks/rsvp_BSibVS972M0)

A Graph Builder based on a file or URL

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_BSibVS972M0

### [Searching through my LinkedIn Portability Data](https://paris.aitinkerers.org/talks/rsvp_EnIzfZNz5wI)

I'm (trying to) build a knowledge database with neo4j referencing the LinkedIn posts I reacted to.

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_EnIzfZNz5wI

### [French Competition Authority (FCA) decisions into graph knowldge](https://paris.aitinkerers.org/talks/rsvp_TF0thCCDMps)

Scrapping and turning +100 pages length decisions of the French Authority into actionable graph knowledge. These decisions are complex domain specific knowledge that instantly fail when used with normal RAG pipelines, therefore graph knowledge is more suitable to capture the complex interactions within these PDF files.

- Event context: Turn Your Knowledge into an API for LLMs - Meetup — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_TF0thCCDMps

### [Introduction to GraphRAG Workshop](https://paris.aitinkerers.org/talks/rsvp_FeCFomaS0Tc)

In this 25-minute workshop, you will learn about GraphRAG, a technique that combines graph databases with generative AI to improve the quality of LLM-generated content. We will explore the terms Retrieval-Augmented Generation (RAG) and Context Engineering, and how GraphRAG can be used in both scenarios. What you will learn What GraphRAG is and how it can improve the quality of LLM-generated content - The common graph shapes used in GraphRAG - How to extract structure from unstructured data and store it in a Knowledge Graph - How to use relationships to provide additional context to vector-based semantic search - How to convert natural language queries into Cypher queries

- Event context: Turn Your Knowledge into an LLM‑Ready API - Hackathon @ Generation AI — 2025-12-09 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_FeCFomaS0Tc

### [Neo4j + Python para análise de redes de colaboração científica](https://saopaulo.aitinkerers.org/talks/rsvp_vsdk71Al65k)

Nesta demo, Pedro Sader Azevedo apresentará como utilizou o Neo4j para transformar dados de coautoria científica em um grafo de colaboração acadêmica, revelando as comunidades ocultas que moldam as redes de pesquisa. Serão abordados os fundamentos da modelagem de dados em grafos, bem como as melhores ferramentas para ciência de dados neste contexto. Além disso, você descobrirá padrões inesperados na vida acadêmica, que revelam o que (ou quem) pode impulsionam seu reconhecimento no mundo da ciência.

- Event context: Ai Tinkerers São Paulo na ESPM com Neo4j e Oracle - Novembro 2025 — 2025-11-27 — São Paulo
- Public talk page: https://saopaulo.aitinkerers.org/talks/rsvp_vsdk71Al65k

### [Anti-cheat com grafos: detectando hackers no call of duty com Neo4j](https://saopaulo.aitinkerers.org/talks/rsvp_NqEKDizKaMA)

Projeto pessoal onde com dados sintéticos crio uma modelagem em grafos usando neo4j e aplico algoritmos de ciência de dados voltados para grafo para identificar hackers, com isso crio um knowledge graph onde aplico graphRAG (LLM + MCP) e tenho como resultado uma "conversa em linguagem natural com meu knowledge graph de anti-cheat"

- Event context: Ai Tinkerers São Paulo e Oracle - Setembro 2025 — 2025-09-25 — São Paulo
- Public talk page: https://saopaulo.aitinkerers.org/talks/rsvp_NqEKDizKaMA

### [LLMs, Grafos e CNPJs: Quem é Sócio de Quem?](https://saopaulo.aitinkerers.org/talks/rsvp_mPc2gZt_myM)

A demonstração apresenta um chatbot construído com LLM e few-shot prompting capaz de traduzir perguntas em linguagem natural para consultas Cypher em um grafo Neo4j contendo a base pública de CNPJs da Receita Federal. Na demo, mostrarei, usando a forma mais simples possivel, como um modelo de linguagem gera consultas a partir do schema.

- Event context: Ai Tinkerers São Paulo com Link School of Business e Oracle - Agosto 2025 — 2025-08-28 — São Paulo
- Public talk page: https://saopaulo.aitinkerers.org/talks/rsvp_mPc2gZt_myM

### [Is AI the Solution to Alert Fatigue? The Problem Can Only Be Solved With a Deep Understanding of How it Was Created.](https://boston.aitinkerers.org/talks/rsvp_bLRjkuOkVhA)

Recent serious work on the problem of false positives and their effects [1] has measured the rate of false positive alerts, in security operations centers, as ranging between 75% to an astounding 99%. Rates in the nineties have been reported by researchers and data scientists doing work on the problem and I observe similar rates in my own measurements across a wide variety of tools and technologies. At the time of this writing, so-called “agentic” or AI-based products promising a solution to alert fatigue are numerous and increasing by the day. Extensive debate is taking place as to the veracity of these claims. Is AI the magical solution to the problem? To understand that, you need to know a little about how we got here, and about how security threat detections are developed, and some mistakes we made along the way. We know a lot about alert fatigue because we actually worked on manufacturing it. Have you ever wondered about how the people who created the problem would approach solving it, if they had a free hand? This is our answer, an open source project named PROTOSTAR that is the work of eight security researchers and data scientists over the past year plus. Is AI the answer? We think it has a place, but not exactly the way it is being sold today. We know a lot about alert fatigue because we actually worked on manufacturing it. Have you ever wondered about how the people who created the problem would approach solving it, if they had a free hand? One of the fascinating things you see, inside software manufacturers, is what goes on during “build weeks.” Twice a year, or on some interval, most teams are free to ignore product schedules and build whatever they choose, often in self-organized teams. What comes out of those periods - tools we build for ourselves vs for demos - tends to look quite different from the products we ship during the other 50 weeks of the year. PROTOSTAR is our approach to alert signal processing, using ML and AI, in a different direction. It is the result of over a years’ work by seven security researchers, developers, and data scientists while on walkabout, we could never have taken the time to do this while at our day jobs. It differs from the current crop of solutions to alert fatigue in several ways; 1) It generates data structures that enable intelligent processing of all detections, by an LLM or by a human, at reasonable cost. 2) is asymptotically efficient, decisioning more than ten thousand alerts per minute, and accuracy increases with volume. 3) It applies ML differently, and thoughtfully, in order to boost signal rather than increase noise. [1] 99% False Positives: A Qualitative Study of SOC Analysts' Perspectives on Security Alarms Authors: Bushra A. Alahmadi, Louise Axon, and Ivan Martinovic, University of Oxford

- Event context: AI Tinkerers Boston: Langchain & Beyond Meetup — 2025-08-25 — Boston
- Public talk page: https://boston.aitinkerers.org/talks/rsvp_bLRjkuOkVhA

### [YouTube Video Analysis Suite](https://dubai.aitinkerers.org/talks/rsvp_8SBsfAl7-Vs)

This session presents a complete walkthrough and live demo of the YouTube Video Analysis Suite, a modular, AI-driven toolkit that transforms any YouTube video into structured, actionable insights. Using MCP (Model Context Protocol), the suite performs advanced tasks like transcript extraction, knowledge graph generation, sentiment and topic analysis, and note creation—all orchestrated through a natural language chat interface and visualized via a Streamlit dashboard. The talk will cover architecture, prompt design, deployment with Docker, and real-world applications in education, research, and content intelligence.

- Event context: AI Tinkerers Dubai Meetup — August 2025 Demo Day — 2025-08-23 — Dubai
- Public talk page: https://dubai.aitinkerers.org/talks/rsvp_8SBsfAl7-Vs

### [Adding graph in your genAI architecture for better and explicable answers](https://paris.aitinkerers.org/talks/rsvp_TklOS1zoJRM)

After explaining what GraphRAG is, where it came from and how it compares to RAG, I will explain the framework that integrates knowledge graphs to enhance the performance of large language models (LLMs). The implementation will explain the four key steps: 1- Knowledge Graph Construction to Extract entities and relationships from input to build a structured knowledge graph. 2 - Vector Search to generate vector embeddings for textual content and perform similarity searches to identify relevant nodes within the graph. 3 - Graph Traversal to navigate through the graph to gather additional context by exploring related nodes and relationships. 4 - Response Generation to utilize the enriched contextual information to produce precise and contextually relevant responses to user queries.

- Event context: AI Tinkerers - Paris Meetup on June 26th - Google Cloud AI — 2025-06-26 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_TklOS1zoJRM

### [Have questions about your building? Why not ask it directly in your language!](https://paris.aitinkerers.org/talks/rsvp_N5ZNYyiu7BE)

A SaaS assistant powered by multi-agent AI, allowing users to interact with building data, urban data, and local urban codes while generating 3D models. This project has been selected by Station F in 2024.

- Event context: AI Tinkerers - Paris Meetup on June 26th - Google Cloud AI — 2025-06-26 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_N5ZNYyiu7BE

### [Group Partner Healthtech & CTO Theodo France](https://paris.aitinkerers.org/talks/rsvp_nQdDHtxpl4Q)

"From Monolith to Modular Architecture: Mapping 500k+ LoC with AI and Graphs" Problem statement : To migrate a complex legacy application (here a monolith in Express to a modular architecture in Hono) we need to map all the existing code to a target architecture (e.g. Controller -> Service -> Model). Because static code analysis alone often fails to capture implicit logic, dynamic routing, or unconventional patterns, manual inspection is typically required—making the mapping process both time-consuming and resource-intensive. By using GenAIScript in combination with an AST parser, we managed to map the whole codebase in 1 hour instead of days, resulting in 6000 nodes that can be visualised with Neo4J.

- Event context: AI Tinkerers - Paris Meetup on June 26th - Google Cloud AI — 2025-06-26 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_nQdDHtxpl4Q

### [Synopse.AI: how can you beat information overload with a knowledge graph?](https://singapore.aitinkerers.org/talks/rsvp_d4qZpzSH3e4)

I'm going to demonstrate how to leverage a recent Neo4J feature (graph storage combined with embedding storage) to create a tailored knowledge graph together with knowledge base taxonomy.

- Event context: AI Tinkerers Singapore: 6th Meetup - April 25th, 2025 — 2025-04-25 — Singapore
- Public talk page: https://singapore.aitinkerers.org/talks/rsvp_d4qZpzSH3e4

### [Translating Financial Data into Qualitative Insights](https://nyc.aitinkerers.org/talks/rsvp_jZXxpn6QCHo)

I will show how you can take ANY csv of financial data and use Ai to create long term structured memory using a knowledge graph that will be used to automate financial analysis. I will show the prompt structure/system and how it is used to create nodes and edges that are highly dense in context and can be retrieve to store tasks.

- Event context: AI Tinkerers NYC Meetup - March 2025 Sponsored by SignalFire and Comet — 2025-03-04 — New York City
- Public talk page: https://nyc.aitinkerers.org/talks/rsvp_jZXxpn6QCHo

### [Never Fill Forms again with "Browser Agents"](https://atlanta.aitinkerers.org/talks/rsvp_PXh82rhmgBw)

I’ll demonstrate a Chrome extension that uses an LLM driven ‘browser agent’ to navigate to a website, locate a job application form, and automatically fill out the fields with high accuracy. During the live demo, I’ll type a command like ‘Navigate to Macy’s careers page and apply for a software engineering job,’ and the browser agent will programmatically follow links, parse the relevant form, and fill out the details in real time. I’ll walk through the core technical components, including some of the code and prompts behind the automated navigation and form-filling logic, and show how the AI integrates with the extension.

- Event context: AI Tinkerers Atlanta February Meetup — 2025-02-27 — Atlanta
- Public talk page: https://atlanta.aitinkerers.org/talks/rsvp_PXh82rhmgBw

### [Airweave: Make Any App Searchable for AI Agents](https://amsterdam.aitinkerers.org/talks/rsvp_53OZ-WKXwj0)

In this talk, we'll show how we built Airweave, an open-source tool that makes source app, API, and database searchable for AI agents. Airweave automates ingestion, transformation, and indexing, while handling authentication, mapping, interweaving, and syncing to vector stores and graph databases. Originally built to solve retrieval challenges in our own AI agent stack, we decided to open-source it. We’ll walk through its architecture, the challenges re ran into when building it (as well as the ones we're facing right now), and structured data retrieval for AI agents.

- Event context: AI Tinkerers Amsterdam - February Edition — 2025-02-26 — Amsterdam
- Public talk page: https://amsterdam.aitinkerers.org/talks/rsvp_53OZ-WKXwj0

### [Talking to a huge product catalog in plain English using LLMs](https://chicago.aitinkerers.org/talks/rsvp_uJ7UIZ0RHQY)

At Fetch, we have a product catalog consisting of millions of products, in extremely granular levels of detail (even more specific than UPCs are!). It can be difficult to easily get summary stats about those products because of the hierarchical structure of the product categories, so I built a knowledge graph on top of the catalog. The cherry on top is a LangChain-driven interface to ask questions like "how many brands do we have in the Red Wine category?" or "what offers are we running right now for snack products?"

- Event context: AI Tinkerers Chicago February Meetup — 2025-02-18 — Chicago
- Public talk page: https://chicago.aitinkerers.org/talks/rsvp_uJ7UIZ0RHQY

## Related Technologies

- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 654 public demos
- [LangChain](https://aitinkerers.org/technologies/langchain) ([Markdown](https://aitinkerers.org/technologies/langchain.md)) — 444 public demos
- [OpenAI API](https://aitinkerers.org/technologies/openai-api) ([Markdown](https://aitinkerers.org/technologies/openai-api.md)) — 517 public demos
- [React](https://aitinkerers.org/technologies/react) ([Markdown](https://aitinkerers.org/technologies/react.md)) — 215 public demos
- [Cypher](https://aitinkerers.org/technologies/cypher) ([Markdown](https://aitinkerers.org/technologies/cypher.md)) — 3 public demos
- [FastAPI](https://aitinkerers.org/technologies/fastapi) ([Markdown](https://aitinkerers.org/technologies/fastapi.md)) — 178 public demos
- [GPT-4](https://aitinkerers.org/technologies/gpt-4) ([Markdown](https://aitinkerers.org/technologies/gpt-4.md)) — 529 public demos
- [Streamlit](https://aitinkerers.org/technologies/streamlit) ([Markdown](https://aitinkerers.org/technologies/streamlit.md)) — 88 public demos
- [ArangoDB](https://aitinkerers.org/technologies/arangodb) ([Markdown](https://aitinkerers.org/technologies/arangodb.md)) — 2 public demos
- [Blazegraph](https://aitinkerers.org/technologies/blazegraph) ([Markdown](https://aitinkerers.org/technologies/blazegraph.md)) — 2 public demos
- [Docker](https://aitinkerers.org/technologies/docker) ([Markdown](https://aitinkerers.org/technologies/docker.md)) — 144 public demos
- [Google Cloud Platform](https://aitinkerers.org/technologies/google-cloud-platform) ([Markdown](https://aitinkerers.org/technologies/google-cloud-platform.md)) — 25 public demos
- [GraphRAG](https://aitinkerers.org/technologies/graphrag) ([Markdown](https://aitinkerers.org/technologies/graphrag.md)) — 13 public demos
- [Llama 3](https://aitinkerers.org/technologies/llama-3) ([Markdown](https://aitinkerers.org/technologies/llama-3.md)) — 38 public demos
- [MCP](https://aitinkerers.org/technologies/mcp) ([Markdown](https://aitinkerers.org/technologies/mcp.md)) — 122 public demos
- [Node](https://aitinkerers.org/technologies/node) ([Markdown](https://aitinkerers.org/technologies/node.md)) — 96 public demos
- [Ollama](https://aitinkerers.org/technologies/ollama) ([Markdown](https://aitinkerers.org/technologies/ollama.md)) — 75 public demos
- [OpenAI](https://aitinkerers.org/technologies/openai) ([Markdown](https://aitinkerers.org/technologies/openai.md)) — 111 public demos

## More Results

- Next: https://aitinkerers.org/technologies/neo4j.md?page=2