Infrastructure should not be
a second job.
Small teams now ship software that would have needed a platform group a decade ago — and they still spend their weeks on brokers, certificates, pool sizes, and cost reviews. Darwa exists to take that work off the roadmap without taking the control away.
The gap is not capability. It is operating cost.
The simple platforms run out of room the moment you need a private network, a replica, or a queue with real retry semantics. The powerful ones hand you a console with four hundred services and a bill you cannot explain. Teams end up paying twice — once in money, once in the attention of the two engineers who understand the setup.
We think the answer is not another abstraction that hides the machine, but a platform that operates it for you and then tells you exactly what it did, what it cost, and what to change. That is why almost every screen in Darwa ends in a specific recommendation rather than a chart.
Six principles we are willing to be held to
Most outages trace back to a setting nobody chose deliberately. We ship the safe default and make the unsafe one explicit.
A platform that fixes something silently teaches you nothing. Ours names the cause, proposes the change, and waits for you.
No per-seat surprises, no metered monitoring, no markup on model tokens. If a number will appear on your invoice, it appears in the UI first.
A free tier that sleeps or expires is a demo. Ours stays awake and keeps its backups, because that is the only way to evaluate honestly.
Brokers, poolers, certificates, cron boxes, image pipelines. Nobody was ever promoted for running them well.
Roadmap items are labelled Soon and nothing above them depends on them. We would rather lose a deal than win one on a feature that does not exist.
Small team, written decisions, on-call together
We are hiring
Platform, product, and support roles across three time zones. The hiring process is four steps and we pay for the take-home exercise.
See open roles