Portfolio

The problem, the build, and what changed.

Screenshots are easy to make look good. These write-ups say what was actually wrong, what we built, and what it measurably did — including the parts that took longer than they should have.

Point of sale · Desktop app

A counter that closes itself out

A document-processing shop ran the register on paper and reconciled it each night against a spreadsheet that never quite matched. Off-the-shelf POS software assumed retail: scan a barcode, take the money, done. Their work is orders with stages, partial payments, and customers who come back three days later.

We built the counter around that instead. Keyboard-first entry so a sale never needs the mouse, order states in colour — unpaid red, partial orange, ready blue, done green — so the whole day reads at a glance, split payments across cash, card and transfer with paid and due calculated rather than typed, real-time search across every order, and a per-day cashier breakdown that turns closing into a glance.

It runs on their own machine with their data on site. No per-terminal subscription, no cloud bill, and no internet required to make a sale.

Desktop Offline-first Local data Split payments
~1 hr
saved nightly at close
$0
recurring software cost
1 day
install to live
Screenshot slot
assets/work/pos-orders.png
Internal tool

Project two — the outcome, in six words

Placeholder. Open with the problem in the client's language — what was being done by hand, how often, and what it cost them in time or mistakes. Don't start with the technology.

Then the build: what you made, and the one or two decisions that mattered. Name the constraint you designed around, because that's what makes a case study read as real work rather than a brochure.

Tag Tag Tag
Screenshot slot
assets/work/project-two.png
Automation

Project three — the outcome, in six words

Placeholder. A shorter entry is fine. Not every project needs three paragraphs — some are one good paragraph and a number.

Tag Tag
00
what improved
00%
by how much
Screenshot slot
assets/work/project-three.png
Toolkit

What we build with.

Chosen for what a small business can actually run and maintain — not for what's fashionable.

01

Applications

  • Python
  • JavaScript / TypeScript
  • Desktop & web front ends
02

Data

  • SQLite & Postgres
  • Local-first storage
  • Migrations & backups
03

Automation

  • Scheduled jobs
  • API integrations
  • Document & file pipelines
04

Applied AI

  • LLM integration
  • Document extraction
  • Guardrails & evaluation

Want something like one of these?

Tell us what your version of the problem looks like and we'll scope it.

Start a project