|
Welcome to This Week in Neo4j, your fix for news from the world of graph databases!
This edition highlights the winning Aura Agents Community Challenge project — PagerDruid, a on-call graph agent that diagnoses microservice incidents using dependency graphs and historical evidence.
We also tackle long-term graph persistence risks, unpack the math behind match-style games with graph algorithms and explore how to build a sovereign, confidential GraphRAG platform that keeps sensitive AI workloads secure in trusted execution environments.
NODES AI, our global graph-and-AI event, is taking place April 15, and the full agenda with themes like Context Graph, GraphRAG, Agents and AI in Production – register now! Road to NODES workshop series has already kicked things off!
Happy Graphing,
Alexander Erdl
|
|
-
Livestream: Discover AuraDB – S02E09 on March 9 & Neo4j Live: Neo4j – The Definitive Guide on March 12
-
Conferences: Find us at AgentCon, New York City on March 9, Gartner D&A, Orlando on March 9-11, HIMSS26, Las Vegas on March 9-12, Data Innovation Summit, Singapore on March 12 & Kubecon Europe, Amsterdam on March 22-24
-
Meetup: Meet us in Pune, IN on March 14, Cologne, DE on March 17, Berlin, DE on March 20, Kochi, IN on March 21 & Mountain View, US on March 24
-
All Neo4j Events: Webinars and More
-
GraphSummit Series: Transform Your Enterprise with Graph and GenAI – Next Stops: Copenhagen on March 10 & Munich on March 17
|
|
Alix de Cremoux is a Machine Learning Engineer at Banque de France, where she develops secure, scalable AI solutions. She has led projects involving the deployment of an intelligent sovereign assistant powered by Mistral LLM and the integration of AI solutions on Azure.
Together with her colleague Gabriel Laffitte, she will be speaking at NODES AI. In their talk “Fusing NLP and Graph: Building a Conversational Agent for Enriched Financial Data “, they will share their experience of building an interactive, context-aware knowledge graph by combining structured corporate data with unstructured financial reports.
|
|
PagerDruid – Blameless On-Call Graph Agent for Microservices As part of our Aura Agents Community Challenge, we had a few fun submissions (check them all out) – the winner is an assistant that diagnoses incidents using graph reasoning over a microservices dependency graph and evidence retrieval from runbooks + incident history. José Luis Latorre recorded a short video to give you an impression of what he built!
If you missed it, don’t worry! In March, we have the
DevCenter Feedback Contest. Help us make the Developer Center better with your feedback and win a cool vintage cap!
The (Very Slowly) Ticking Time-Bomb in Your Graph Persistence Stack
Jasper Blues highlights a “time bomb” in graph persistence stacks: how seemingly benign choices in storage layers, backup strategies and schema evolution can silently degrade performance or data integrity over time. The article outlines practical pitfalls (e.g. unbounded relationship growth, inefficient indexing and schema drift). It offers tactics to future-proof your graph database, ensuring consistent performance and robustness as data scales.
The Math and the Graph Behind a Popular Match Game
This post by Jeremy Adams breaks down the graph logic and mathematics behind a popular match-style game, showing how swaps, matches and cascading clears can be modelled as graph traversal and pattern recognition problems. By representing game pieces and adjacency relationships in Neo4j, developers can use Cypher and graph algorithms to detect match patterns, optimise scoring logic, and reason about game state transitions more intuitively.
Building a Highly Secure and Sovereign GraphRAG Platform with Confidential Computing
Marco De Luca explores how to build a highly secure, sovereign GraphRAG platform by combining Neo4j’s graph intelligence stack with confidential computing capabilities – protecting sensitive data and AI workloads while in use. By running GraphRAG pipelines within confidential computing environments, you can ensure data remains secure during processing and meet stringent enterprise or regulatory requirements without exposing plaintext to cloud operators or external systems – a critical step toward enterprise-grade, privacy-preserving AI systems.
|
|
|
Please share it if you like it!
|
|
|
|
