Bring me the messy problem.
I turn vague ideas into working tools, fast enough that you can test them, use them, or kill them with confidence. Fifteen years of shipping for the web, the last three building with AI every day.
Start small, then build only what survives contact with reality.
Start here
Discovery + Prototype Plan
€400–€500
A focused diagnostic for a messy idea, so it gets a practical technical direction before anyone wastes more time on it.
- One deep-dive call
- Problem breakdown and bottleneck analysis
- Recommended tools and architecture
- Prototype scope and a next-step plan
Then build
Prototype Sprint
from €2k–€3k
A short, fixed-scope build that turns the idea into something real enough to test, show, or learn from.
- A working demo or proof of concept
- Fast iteration over a tight time window
- Clear scope instead of open-ended consulting drift
- Handoff notes and realistic next steps
For your team
Practical AI Workshop
from €750
A hands-on session for your team on today's AI tools: what they do well, where they fail, and what is worth adopting.
- Walkthroughs of real tools and workflows
- Your use cases, discussed honestly
- Recommended adoption paths
- No hype
How it goes
- 1
You write a few lines
What's stuck, who it's for, and what you've tried. Messy is fine.
- 2
We talk for an hour
I ask a lot of questions. By the end we both know if this is worth building.
- 3
I build the smallest real thing
Not a slide deck. Something you can click, run, or put in front of users.
- 4
You decide
Keep it, change it, or kill it. All three are good outcomes when they're made on evidence.
Things I've built like this
All public. Look before you write.
- ai-pricing
A live tool comparing AI models on price and performance. I built it because I needed the answer myself.
- BasisExamenOnline
A whole product with one partner: learner app, teacher admin, payments, all of it running in production.
- Mochi
An AI agent that reads mail, calendar and chats through narrow tools and acts only after a yes. The guardrails are the product.
- tg-archive
Self-hosted sync, search and storage for years of messages. The unglamorous systems side.
Where this works best
- Agency owners who need internal tools, automation, or fast prototypes
- Ops-heavy teams drowning in manual work or messy workflows
- Founders and operators with good ideas and no one technical enough to test them quickly
- Teams that know AI matters and want practical guidance instead of trend-parroting
What I avoid
- Committee-heavy enterprise procurement
- Vague unpaid brainstorming
- Tiny favour-sized custom work with fuzzy scope
- Long strategy decks with nothing built