Rotterdam in miniature

thijs.md

The quay

I’m Thijs.

I build AI systems and software from Rotterdam, after twelve years in data and AI at companies including KPN, Eneco and Free Now. At Free Now in Hamburg I was Head of Data.

When you talk to me, you talk to the person who builds it. If a job outgrows one person, I bring in one senior I have known for years.

Rotterdam On the homepage I sit by the Maas too, with the same laptop.

More about my role

approach.md

Erasmus Bridge

From pilot to production.

A working demo is not yet a system your team uses every day. That is why I work in three steps, and the last one is the handover.

Establish the problem first. We choose one problem and agree how you will know it has been solved.
Test against your everyday work, with your own data.
Agree who takes it forward: who maintains it, who steps in when something goes wrong and how you check the next version.

Rotterdam The Erasmus Bridge opened in 1996. It is 802 metres long with a 139-metre pylon. Locals call it the Swan.

How we work together, including scope and pricing

company-brain.md

Central Library

A company brain.

A company brain brings knowledge from documents, notes and decisions together in one place that people can search with AI, with the source next to every answer.

I build it with the team, test it on their real questions and help get it into daily work.

Rotterdam The Central Library opened in 1983. The yellow pipes on its facade are ventilation ducts.

Who I build this for

data.md

The port

Data & analytics.

Bringing data from different sources together, checking that the numbers add up and turning them into analyses and reports for questions from other departments.

With the team I first pin down which questions the numbers must answer and how each figure is calculated. Then I build the view.

Rotterdam The Port of Rotterdam is the largest port in Europe.

Clients and partners

software.md

Building site on Zuid

Software development.

From a clearly defined question to a working application. I assess which AI applications are worth it and then build the software people use. Advice and technical work sit with the same person.

Examples of such jobs: preparing a meeting, spotting a price change at a competitor and selecting prospects. They are on the homepage as demos, with made-up data.

Rotterdam De Rotterdam, the three towers nearby, was designed by OMA and completed in 2013.

See the examples

ai-training.md

Van Nelle Factory

AI training.

I guide people in working with AI. In hands-on sessions they use AI for their own tasks and learn when to check an output before building on it.

Rotterdam The Van Nelle Factory has been a UNESCO World Heritage Site since 2014.

Clients and partners

products.md

Markthal

Also built by me.

Three products I built that are running today.

Signal Match turns a candidate CV into a client-ready profile in the agency’s house style, with anonymisation built in. Running at recruitment agencies since May 2026.
Marketingburos is an independent database of Dutch marketing agencies, with a wizard that builds a shortlist.
Business Boosters finds B2B prospects for targeted outreach. Built for a client, who owns the brand and the product.

Rotterdam The Markthal ceiling is a single artwork: the Horn of Plenty by Arno Coenen.

memo.md

Euromast

The view from 35+ AI projects.

Since February 2025 I have done more than 35 AI implementations, and not every one went well. My memo has fourteen lessons in three phases: before you sign, while building and after go-live. Three of them:

Without a baseline, every result is an opinion.
Internal green is not proof. Client data is the truth, and the test should belong to the client.
Never leave an agent behind without an owner.

Rotterdam The Euromast is 185 metres tall. The rotating Euroscoop takes you to the top.

Lessons from 35+ AI projects

done.md

You have seen all of Rotterdam.

Eight places, one person who builds it. Tell me where your team loses time or lacks information. After the call you know whether I would build it and what the first step is.

Let’s talk Lessons from 35+ AI projects

30 minutes · free, no obligation

find-the-answer.md

Interactive example · fictional data. This example runs locally in your browser. Nothing is sent.

Find the answer.

“What do I need to know before the next project meeting?”

You have a project meeting soon. What you need to know is spread over five places: minutes, email, the shared drive, chat and the CRM.

Round 1: steer the water taxi past the jetties and collect the three sources that matter. Round 2: the same, with a company brain.

Steer with the arrow keys or WASD. With a mouse or finger: press and hold where you want to go.

round-1.md

Found.

It took you 0:00 and 5 places. Now the same question, with a company brain.

briefing.md

Draft for your review

Your meeting, prepared.

Decision changedThe pilot starts with support. Sales comes later. This replaces the earlier plan to start with both teams.Minutes [1]Email [2]
Next commitmentSupport will provide example questions on the coming Friday. Use these to discuss what the first version needs to answer.Shared drive [3]
Still openWho will own the knowledge base? The sources do not answer this yet.Shared drive [3]

Check the changed decision and agree who will maintain the knowledge base.

result.md

Same question, different afternoon.

Searching yourself: 0:00
With a company brain: 0:00

That is what a company brain does: knowledge from documents, notes and decisions in one place, with the source next to every answer. And when the answer is not there, it says so. The judgment stays with you.

Let’s talk Lessons from 35+ AI projects

30 minutes · free, no obligation