Three steps, each with an end.
Every project starts small and fixed, so you know what you're paying for and when it's done. I take on a few projects a year, so I'll tell you up front whether yours fits.
Discovery sprint
We work out what to build first and how. By the end you have something useful whether or not we keep working together.
- A scoped roadmap: what's in the first version, what waits, and why
- An architecture sketch: the stack, the data, and the hard parts
- A clickable prototype, or a technical spike on the riskiest piece
- A fixed-price quote for the build
Build
I build what the sprint scoped, usually over six to ten weeks. You see working software every week, not a status report. Changes to scope are fine; they get written down and priced before they happen.
- First versions: TypeScript, React and Postgres, deployed and monitored
- AI features: LLM features in your existing product, with evaluations so you can tell whether they work
After launch
A set number of days each month for fixes, small features and keeping dependencies current. Or nothing at all: the code, the notes and the deploy are yours, and documented well enough for someone else to pick up.
Probably not a fit
- Marketing sites and CMS work. There are people who do it better and for less.
- White-label work for agencies.
- Hourly staff augmentation with no defined scope.
Start with a short form.
A short form, about a minute. I read every one myself and reply by email, usually within two business days.
Start a project