I work out which numbers matter to a company, then build the software that improves them.
That starts with how the company actually makes money, and how finance and the teams fit together to produce it. Then comes the building: more sales, lower costs, or less work done by hand.
I work on my own and I do the building myself. Mostly for founders who need something made, and for growing companies that want to use AI but have nobody to do it yet.
Six kinds of work, one person doing them.
Services in full →Building the software
The thing itself: an assistant that does a job, a system that sets prices, a tool your team opens every day.
Working out what to build
Which ideas are worth building, in what order, and which ones to drop. Usually a short piece of work before anything gets made.
Connecting AI to your own systems
So your team can ask a question in plain language and get an answer from your own numbers.
Dashboards and reporting
Reports people actually trust, where a number means the same thing everywhere in the company.
Getting your data in order
Collecting data from the places it lives, cleaning it up, and keeping it current.
Helping your team use AI well
Getting your developers genuinely faster with AI coding tools, instead of just having them installed.
Miinto
An online fashion marketplace selling in twelve European countries, where I lead business intelligence. The tools below are ones I built and run in-house: pricing, forecasting, reporting, and the shared data everything else reads from.
Eight tools in daily use, each one built from end to end. It is the closest thing to a portfolio I have, and the reason I know this kind of system holds up once real people depend on it.
Data first, then AI.
Six years in online fashion and marketplaces — data scientist at Stylepit, head of BI and data science at The Vintage Bar, and head of business intelligence at Miinto since 2023. I spent those years on the unglamorous half: getting data collected, cleaned and trusted. That is where most AI projects actually go wrong, so it is where I start.
Enumstudio is my own practice, open since August 2026. The full history is on LinkedIn.
I also build and run my own products under Enumlabs: a shared room where several AI agents working on the same code can keep each other informed, and a way to search the books you own and get the exact pages back. The AI tools I set up for clients are ones I use on myself first.
Tell me what is slow, done by hand, or nobody trusts.
Either a project with a clear goal, or a monthly arrangement where I act as your AI and data team. Either way you get a straight answer on whether it is worth building at all.