About
Long-running ideas about structure in data, built so you can try them.
Patternode develops a set of ideas about structure in data: how irregular series, images and knowledge can all be read as graphs of patterns. Each idea is written up, and where it has got far enough, there is a lab you can open.
Founder · New Zealand
An enterprise and solution architect working on AI. Dermot founded Patternode and leads its research, from the mathematics of shard trees to the knowledge graphs and the labs.
What the work draws on
- Mathematical modellingShard trees
- Ontology designEarnings Lab
- Agent evaluationEarnings Lab
- Search at any scaleShape search
- Enterprise architectureThe method
- Cloud and TerraformEvery lab
- Data visualisationEvery lab
An AI-native way of working
Architecture, research and delivery run as files that AI agents read, check and progress, under a method Patternode publishes as open source. It lets a lean team carry 10 research lines and several working labs at once.
- 10research ideas on the map, each written up
- 9open-source repositories, including the method itself
- 8labs, from live demos to work in progress
Get in touch
If an idea here overlaps with something you are working on, or you would like a walkthrough of a lab, we would be glad to hear from you. To ask your own questions in the Earnings Lab chat, request access.
Patternode Limited · NZBN 9429052081224 · Company number 9245509 · New Zealand