Installation, 3 days€ 4.850,00
Materials€ 2.310,00
VAT · 6% renovationauto
Total incl.€ 7.589,60
Portfolio
Open for new work
I build the systems companies actually run on.
- Leads
- Quotes
- Site
- Invoices
- Portal
Custom CRM and operations platforms. From the first inbound lead to the settled invoice — every role gets an interface built for the work it does, rather than one screen with everything hidden behind permissions.
Stack
- TypeScript
- React
- Postgres
- Supabase
- Swift
four systems, not a gallery
- Site · 14 points open Project-Point Site registration and project management for the building trade — “Geen enkel punt raakt kwijt”. My own product: landing page, app and customer portal.
Open
Closed
Closed
- Case · 6 roles · week 11 Operations platform Sales, planning, field execution, logistics, finance and a customer portal on one database — the system an entire installation company spends its day inside, from the first lead to the reconciled payment.
- Signup flow · code extracted catchbox Disposable inboxes for developers — one mailbox per flow, so signup, billing and password resets never land in the same pile. It hands you the one-time code and the action link, and reads the headers for why a mail went to spam. Node, no dependencies.
$ catchbox code --wait
482910
copied
Link
- macOS 27 · menu bar Stash A menu bar hider for macOS 27. The rewrite drew the menu bar as a single window and broke every tool in that category at once — the repository documents what changed, measured on a real machine rather than guessed at.
Method
How I work
four rules, paid for in incidents
-
Correctness before cleverness wherever money is involved
Line-item pricing, VAT regimes and invoice totals are guarded by database-level invariant checks that run on a schedule and alert only when the state changes. An alarm that repeats every fifteen minutes gets dismissed, and is worthless from then on.
- In practice
- Every money rule is a constraint or a scheduled check, never a comment.
- Without it
- Totals that drift apart quietly, and a customer who finds it before you do.
-
Access control is not something you bolt on afterwards
A public customer portal sitting on the same database as internal finance data means every row-level policy and every grant is load-bearing. I assume a policy is broken until I have watched it refuse the request.
- In practice
- New endpoint, new policy — and a test that signs in as the wrong user first.
- Without it
- One missing predicate turns a portal into an export of everything.
-
Every incident becomes a written rule
Anything that cost real time gets documented in the repository — the rule, and the story behind it. A warning in prose only works if you happen to read it right before you would have made the mistake, so the ones that matter become automated checks instead.
- In practice
- The rule ships with the fix, in the same commit, with the story attached.
- Without it
- The same outage twice, six months apart, by someone who never heard about the first.
-
Measured, not assumed
Before I claim something works, I run the check and read the output. That habit is the single biggest difference between a system you trust and one you hope about.
- In practice
- No “should work”. Run it, read it, quote the line that proves it.
- Without it
- A deploy that was green in your head and red in production.
Run the check, read the output, then say it works.