We come in where failure isn’t an option: customs invoicing, identity verification in banking, state document issuance, a HIPAA meetings platform, a construction ERP, and a US marketplace where we have held technical leadership for four years. Two of them were acquired as products.
the minimum cost of poor software quality in the US economy.CISQ, 2022
Problem, constraint, measured outcome. Every page starts with what could not change.
“MySQL was the only thing that could not move.”
“Web and mobile must not drift apart.”
“Neither the queries nor the schema could change.”
“Nothing was broken in a way anyone would notice.”
The difference between ugly and expensive: every change to a surface someone relies on is paid for by someone else. We stabilize first and rewrite only when it is earned.
Four pieces of evidence for the same judgement, across three client systems.
“No single database does all three jobs.”
“A convention with no owner is a suggestion.”
Alongside client work we build operational software for small companies in the physical world — the ones with vehicles, people in the field and paperwork that has to be right. Three layers of the same idea.
Profitability tracking for small companies, entered through VAT preparation for their accountant. The data the accountant enters anyway fills the reports, so the company never starts from an empty screen.
Stands on an ERP with inventory, suppliers and invoicing (2025).
Identity verification as infrastructure — orchestration and user consent on top of existing vendor services.
A continuation of a KYC system for a financial institution (2020) and a state document issuance system (2024) — both ends of the identity chain.
Fleet management and telematics for rental companies, with vehicle data available for the duration of a rental. Identity OS is its first internal consumer.
Developed with our own fleet as the test environment; five fleet and driver dashboards built 2022–26.
None of them has a user other than us yet. That changes in the first quarter.
Six years building systems that can’t stop, from Sarajevo for clients in the US, Germany and Sweden. Eighteen case studies, two products sold, four years of technical leadership.
We don’t write components to a specification. We come in where the constraint is already set — the schema can’t be touched, the queries aren’t ours, the old app can’t be switched off — and look for what can change underneath it.
Full background →A single 30-minute conversation, no obligation, is enough to know whether we fit. We reply within one business day.
Start a conversation