The machine: custom software, automations and intelligent agents that work 24/7 — no fatigue, no errors. Engineering built on an M.Sc. in Business and graduate computer science studies at Aalto University.
Beyond automation, I replace SaaS products that will not bend to your needs — with custom-built software.
Mirror Line's routing tool — the industry's old gold standard — aged out, with no more updates coming. I built an equivalent map-based routing application, tailored: it reads data straight from their ERP and bends to their exact process.
One of 12 automations for a bank with a €7bn balance sheet: a recurring manual step replaced by a system that runs without human hands.
A bot monitors competitor prices and sales volumes around the clock — so projects are priced right and built in the right locations.
The bot reacts to a new lead in under half a second — contact goes out before competitors even open their email. Delivered for an international client.
Stock levels, replenishment and order flow run on a blend of AI and automation — manual routine work eliminated.
I have built 12 custom automations for a bank with a €7 billion balance sheet — combined savings of 1,350 working hours every year. That is proof of what automation can do in a large, regulated, error-critical environment — and the same engineering discipline goes into every Nordlake delivery, whatever the size of your company.
Every engagement is different — but the pattern usually goes like this:
A short call or an email exchange: tell me your situation, and we'll see if this makes sense. No strings attached.
If the pain point is clear, you get a proposal right away. If you're not sure where to start, an AI audit maps your operations and pinpoints the best targets — with a proposal of its own.
I build the solution all the way to production — and the results are measured in hours and euros.
Code is the tool. Your bottom line is the goal.
The generalist's toolkit: business (M.Sc.), engineering (graduate studies at Aalto), programming, machine learning, AI. Your problems don't respect skill boundaries — so neither should the person solving them.
In 2018, Roope Keto was sitting at Ernst & Young over IT-audit workpapers when he noticed he was doing the same steps to the same formula, week after week. Couldn't a computer do this? Neural networks were reaching a new level of capability at the time. He completed machine learning studies, began graduate studies in computer science at Aalto University, and moved to Robonomist as a data scientist.
The foundations were laid earlier: programming, physics and mathematics in high school — and a Ministry of Education scholarship that took him to an astrophysics summer school in Moscow. Then an M.Sc. in Business from the University of Tampere, majoring in accounting, with three years of bookkeeping for a welding company alongside his studies.
The Robonomist years put into production, among other things, the data visualisations for Business Finland's export handbook and a housing-market application with 50+ visualisations, including a machine-learning model that forecast housing prices. Since 2023 he has built twelve automations for a bank with a €7 billion balance sheet, saving 1,350 working hours a year. In 2026 he founded Nordlake around that same question.
Off the clock, Keto — a former national-level javelin thrower — spends his time with his family in Ostrobothnia.
I don't sell marketing hype or off-the-shelf ChatGPT prompts that break at the first system update. I keep things fully transparent — my complete work history and technical skills are right there in the CV:
Tell me where your working hours leak — or which software needs replacing. You talk to the person who builds it — we have no salespeople.
Reply within one business day.