Hi, I’m Willy.
I build AI systems that do real work, and I help teams figure out which ones are worth building.
[FILL IN: background. Where you’ve worked, what you’ve built and shipped, and the moment you realised most AI projects fail for boring reasons rather than technical ones.]
[FILL IN: why Tomtar Labs. Why independent, who you like working with and what you won’t do.]
[FILL IN: how you work. Short, direct, hands-on, in the code. About 150–250 words in total.]
The toolbox
- Claude
- OpenAI
- Gemini
- Open-weight models
- Python
- TypeScript
- SQL
- Vercel
- Next.js
- Postgres
- [Cloud]
- Evals
- RAG
- Agents + tool use
- Observability
Why Tomtar?
Tomtar are the tomtes of Scandinavian folklore: small household spirits who quietly keep the farm running overnight. Treat them well and the work gets done. That’s the job good AI automation should do: helpful, mostly invisible, and making everything else run.
Four rules I don’t bend.
- 01
Ship small and early
You see working software in the first weeks, not a slide at the end.
- 02
Measure everything
Every system ships with a baseline and a way to tell if it’s working.
- 03
Your data stays yours
Built in your accounts where possible, with providers that don’t train on it.
- 04
No lock-in
You own the code, the prompts and the docs. Your team can run it without me.