About us
Four years labelling data for computer vision teams taught us exactly where AI projects break. Now we build the models and the products too - for clients who have kept us on for years.
What we do
We take on the whole span of an AI product: the data pipeline that feeds a model, the model itself, the service that exposes it, and the interface where someone acts on its output.
That scope exists because splitting it across teams is where projects die. A model handed over without a deployment path is a research artefact, not a system.
How we work
Since 2021 we have worked on computer vision and machine learning problems where the difficulty is in the data, not the architecture. Millions of images across live projects for teams in the UK, US, Europe, the Middle East and beyond.
That gave us an unusually direct view of why models underperform, and it is why clients keep extending scope: first the model, then the service, then the product around it. Our longest engagement has run continuously for over three years, with a dedicated team of 30 on that one account.
Our journey
Set up to build computer vision systems for teams that needed models working on their own imagery, not on a benchmark.
Sorted Tech in the UK and MiniAi Live in the US both started with us this year. Sorted Tech is still with us and remains our longest engagement; the MiniAi Live work ran for three years. Scope widened from vision into broader engineering.
Engagements widened from single models to complete development stacks, with dedicated teams owning labelling, training, pipelines, deployment and the product surface.
Completed a full PPE compliance platform in Canada - dataset through to edge deployment and portal - alongside agent-based automation for port and logistics operations, and a new engagement in Spain.
Team
A team of 30+ across engineering, modelling, and data work. These are the people who scope your project and stay accountable for it.
Chief Executive Officer
Runs the business and client relationships. Focused on keeping engagements long-term rather than transactional.
Chief Technology Officer
Owns technical direction across AI and product engineering, and is accountable for whether a model is good enough to ship.
Life here
We work in person. It matters for the kind of problems we take on, where the useful conversation is usually the unplanned one.
