Hey everyone,
I’m thrilled to share my latest project— Relational Graph Neural Networks (RGNNs)—which I believe could change the way we interpret structured data and reshape how we think about graph databases in AI.
Traditional graph databases have always been powerful for modeling relationships between entities, but I’ve been working on taking this to the next level. RGNNs treat each node as more than just an isolated point; think of nodes as tables with attributes that allow for a richer contextual understanding of the data. Instead of just connecting nodes, RGNNs allow the use of policies that govern how nodes interact, enabling the creation of additional or even different links depending on the context.
This means we can interpret complex, multi-domain data with real-time flexibility, moving beyond static relationships to something more dynamic, adaptable, and closer to how we reason as humans. It’s a step toward true contextual AI and could redefine how AI systems handle everything from legal frameworks to financial systems and healthcare data.
But here’s the kicker: I’ve hit my API limits, and as much as I’d love to keep pushing forward, I’m currently broke! 
So, I’m reaching out to the community for collaborators who can help test this framework. If you have the resources, curiosity, and passion for exploring new frontiers in AI, I would love to work together and build something game-changing.
What I need from you:
• Testing: Try running the RGNN models in real-world scenarios to see how well they handle structured data.
• Feedback: Provide input on the framework, point out areas for improvement, or suggest new use cases.
• Collaboration: Whether you’re into machine learning, data science, or software development, your skills could help take this idea to the next level.
This is an exciting frontier, and with the right collaboration, I’m confident we can revolutionize how AI interprets and understands structured data.
Feel free to reach out if you’re interested in collaborating or want to chat about the project.
Callum Maystone
Email: [my email]
Medium: Medium
LinkedIn: https://www.linkedin.com/in/callum-maystone-57b00932/