Engineering
We take on bounded technical mandates for teams with a concrete problem: one place in the system that is too slow, too expensive or too fragile. No body leasing, no open-ended timesheet. One goal, one period, one result you can measure again yourself.
A service that folds under load, a database that becomes the bottleneck, a network path that eats latency. We measure where the time really goes and fix the three most expensive places. Where it pays, we rewrite the path in Rust; where it does not, we say so.
Clusters that need more attention than the applications on them. We look at etcd, the network stack, the runtime and the deploy path, and bring operations to a state a small team can keep.
Away from a provider, onto a new platform, out of a monolith or back into one. Planned so that operations continue during the move and the old state stays reachable at any time.
OpenID Connect, SAML for your enterprise customers, passkeys, transactional mail that gets delivered. On our platform or yours; we are attached to neither.
The fixed-price audit
Five days. We measure your system under real load, find the three slowest places and deliver a report with numbers, causes and the effort to fix each one. You then decide whether to do it yourself, engage us or do nothing. The report is yours.
What we do not do: staff teams by the hour, projects without a defined end, and promises of availability for systems we do not run ourselves. We name prices in the call, because they depend on the problem, not on the hour.