We're a young studio, and we won't pretend otherwise. As client projects close, and clients agree to be named, they appear here as case studies with verified numbers. Until then, these work notes show exactly how we think and build.
A circular economy project with Wargön Innovation, funded by Vinnova: image recognition and near infrared spectroscopy fused into an automated sorting process that sorts toward real buyer demand.
How we built an AI talent matching platform for consulting firms and recruiters, now in production at Itmotive and Braha IT. What we built, what we ruled out, and where we stopped the AI.
One traceable workspace for compliance assessment across NIS2, the EU AI Act, ISO 27001 and DPIA. Where the AI sits, where it deliberately does not, and what is still being built.
We launched an app that valued used clothes from a photo and matched them to local buyers. Then a regulation reshaped the market. What the technology proved, and what we misread.
The five checks we run to find out how exposed a stack actually is: weights, logs, jurisdiction, terms drift, and exit cost. A checklist you can run on your own.
The stack we reach for when a system has to run on-prem for a decade: open-weight models, portable serving, and an exit path at every layer.
Our checklist for taking a model from a GPU workstation to constrained hardware: quantisation, latency budgets, and what breaks on the way.
Risk classification, data lineage, model documentation: how we translate regulation into backlog items a team can actually ship.
A project earns a place on this page only when it can carry a name and a number. Every case study we publish follows the same four-part structure, and until a project can be told that way, it stays a work note.
The fastest route to a case study is a project that works. Start with a two-week ML Feasibility Sprint (fixed scope, från 95 000 kr), and we'll earn the right to write it up.