SW
Where we work
Two industries where "move fast" is the wrong advice
In finance and healthcare the cost of being wrong is not a bad quarter. It is money that moved to the wrong account, a record disclosed to the wrong person, a regulator asking for evidence you cannot produce.
That does not mean building slowly. It means building so that changes are small, reversible and evidenced — which, once it is in place, is faster than the alternative, because nobody is frightened of the deploy button.
What we build
Four kinds of work
Breaking up what has grown together
Carving a large application into parts along real business boundaries, so one change stops meaning one full regression cycle and one release train.
Making systems talk
Payment rails, core banking, HL7 and FHIR, insurers, labs, national services — and the ones with no API at all, which need patience rather than cleverness.
New products and internal tools
Web applications, APIs and the unglamorous internal software that a department actually runs on and nobody ever budgets for.
Projects that have stalled
Half-finished, undocumented, or written by people who have left. We start by reading it, not by proposing a rewrite.
How we work
Rules that come from the industries, not from us
PCI DSS, GDPR, medical device software rules, whatever your regulator expects of change control — these shape the design from the first week rather than being bolted on before an audit. In practice that means narrowing what handles sensitive data, so most of the system falls outside the expensive scope entirely.
Every release leaves evidence behind: what changed, who approved it, what was tested, how to put it back. Not because it is satisfying, but because the day somebody asks, the answer either exists or it does not.
How it runs
An engagement, start to finish
Read what exists
Code, data, integrations and the parts only one person understands. No recommendations until the picture is real.
Agree the first thing worth shipping
Small enough to be in production in weeks, useful enough that stopping there would still have been worth it.
Build in the open
Your repositories, your infrastructure, your team in the reviews. Nothing happens in a room you cannot see into.
Release without ceremony
Automated tests, staged rollout, a way back. If a deploy needs a meeting, the pipeline is wrong.
Leave it maintainable
Documented, tested and handed to your engineers. If nobody on your side can change it after we go, we have failed at the actual job.
Working with us
Small teams, your codebase
We work in small teams alongside yours rather than taking a specification away for six months. Where you have engineers, they stay on the work and finish knowing the system; where you do not, we build so that a future team can pick it up without a rewrite.
We would rather scope a project down until it is honest than win it and discover in month four that the estimate was fiction.