Pierre Halftermeyer

[Pierre Halftermeyer @halftep]

Looks like #AdventOfCode 2024 is #AdventOfGraph!
Today - day 6 : path finding in a maze with a twist! Like always, if you get the #GraphModeling part right, queries are easy to ✍️ and 🚀efficient!
Here ➡️#Cypher with #QuantifiedPathPatterns for the win! #Neo4j

Puzzle: https://t.co/L4figcLzck
Solution: https://t.co/Wisxn53LcJ

[Emil Eifrem @emileifrem]

Folks, today we're announcing the acquisition of @graph_aware, moving up to the stack, and taking on black-box intelligence tools like Palantir Gotham. Let's go! 🚀

For 10+ years GraphAware has been our partner. They're a leader in mission-critical intelligence analysis. Together, as one organization, we're building an open-standards alternative to black-box intelligence tools.

1️⃣ The database belongs to them
2️⃣ The data belongs to them
3️⃣ Every choice they make with it belongs to them

Nothing important is hidden behind proprietary logic. They can see how data moves, understand how results are produced, and trust the system they rely on. This is not theory, this is not a vision, this is not hypothetical. What's so wonderful about GraphAware Hume is that it's already deployed by the Western Australia Police, the European Commission and government agencies across the US, UK, EU, and Australia.

Looking ahead, we're building the best possible open, AI-powered intelligence platform that agencies can deploy on their own terms, own their data on, and build on for years to come. One open-standards stack. Deeper querying. Better performance at scale. A tighter integration between analysis and data.

Welcome, GraphAware!

[Pierre Halftermeyer @halftep]

@HawkEmDownChris Abu Diaby

[Neo4j @neo4j]

Neo4j has always used synchronous I/O. Every read or write block the thread until the OS finishes. Simple, predictable, and a real bottleneck in cloud environments where storage latency is higher.
2026.04 changes that. Async I/O via io_uring is now available for the background page evictor and checkpointer on Linux, meaning Neo4j can issue requests without waiting and keep working while storage catches up.
It's Enterprise only, requires liburing and JDK 25+, and is off by default. One config line to turn it on: server.memory.pagecache.async=true
more:
https://t.co/xjggbqO0Fp

[Pierre Halftermeyer @halftep]

@georgeeddy2 https://t.co/bAhjYRurcb

[Neo4j @neo4j]

Going to #GoogleCloudNext?

1️⃣ Visit our booth #2717 to see cool demos and talk to our experts about how graphs can ground your #LLMs with accurate context
2️⃣ Request a meeting with our specialist to get your data AI-ready
3️⃣ On April 23, we are hosting an executive dinner together with @Databricks - make sure you request to attend!

@googlecloud @gcloudpartners

https://t.co/rgPt8BMgXq

[Neo4j @neo4j]

Neo4j Live: Neo4j - The Definitive Guide https://t.co/SUhWgCMhwF

[Pierre Halftermeyer @halftep]

@thekerukeion @HappyPunch @FFLose

[Pierre Halftermeyer @halftep]

@GeorgesDuroyJr Markdown + Marp

[Neo4j @neo4j]

Announcement 🎊 GRAPH TYPE is now in Public Preview in #Neo4j 2026.02!

What this means:

✓ Strict data quality - Enforce property types, mandatory fields, label implications, and relationship endpoints in one place
✓ Flexible governance - Validate your core model without sacrificing Neo4j's schema-flexibility for ad-hoc evolution
✓ Simple lifecycle management - Four commands handle everything: SET, ADD, ALTER, and DROP

Read more: https://t.co/tlzmy8TleK

[Neo4j @neo4j]

Building AI-ready systems starts with the right data foundation.

Join Jesús Barrasa for a webinar and learn how domain modeling and ontology integration strengthen graph schema design by embedding meaning, context, and shared understanding directly into your knowledge graph.

You’ll discover:
• How domain modeling and ontologies create scalable, semantically rich data foundations
• Why a formal semantic layer improves consistency, clarity, and interoperability
• How #AI can accelerate modeling, enhance semantic alignment, and support schema evolution

Whether you’re designing a new graph or evolving an existing one, this webinar will help you build connected data models that are truly AI-ready.

Asia Pacific: https://t.co/iCEukALQF6

Europe: https://t.co/RgnRRJYpjn

Americas: https://t.co/WWH7QRIzZJ

[Pierre Halftermeyer @halftep]

@BrianRoemmele Just put random people in this rocket to live the French 70's experience. https://t.co/dZ0mJpXfhD

[Pierre Halftermeyer @halftep]

@90sfootball Dugarry’s header 😛

[Pierre Halftermeyer @halftep]

@jkoonyango @90sfootball Ballon d'Or was for European players only until 1995.

[Pierre Halftermeyer @halftep]

@Bruno_Constant Haaland aurait dû tomber pour réclamer la faute

[Neo4j @neo4j]

🔍 Neo4j v2026.01 introduces #Vector search with filters as a preview feature: you can apply predicates inside the vector index at query time so the index behaves like it only contains vectors that match your criteria.
Helping with:
-latency low
-relevant results

https://t.co/cbXvmEbXKa

[Pierre Halftermeyer @halftep]

@WilooFootball JPP

[Pierre Halftermeyer @halftep]

@staderochelais 🙏💪

[Neo4j @neo4j]

Insights at scale with Infinigraph 🚀

We are thrilled to announce that #Infinigraph is now generally available and helping companies with billions of data points, allowing:

👉 Building Massive Knowledge Graphs.
👉 Supporting Agentic AI: Provide AI agents with "long-term memory" that grows with your business.
👉 Giving real-time context: Deliver structured context to #LLMs in milliseconds.

Need to scale? Read how Infinigraph can help.

https://t.co/C4YWQXGNMV

#Neo4j

[Pierre Halftermeyer @halftep]

@PerseusLeGrand “a minima” pour “au moins”.
Pédant + fautif => irritant.

[Pierre Halftermeyer @halftep]

@90sfootball Klinsmann