Early architecture is the most expensive thing to revisit later
A venture backed platform business had a team too small to justify a full time chief technology officer. It was also making the decisions that most need one: how the platform is put together, what it is built on, and which engineers to hire first. Those choices get made quickly, by whoever is in the room, and then they become structural.
Revisiting them later is expensive, because by then the product, the data and the team all sit on top of the original choice. Reviewing an architecture is one kind of job. Setting one, and then hiring the people who will live with it, is a seat rather than a piece of advice.
We put senior judgment where the decisions actually were
We scoped the engagement around architecture and hiring rather than around day to day delivery, then sourced former chief technology officers, engineering directors and platform architects who had built comparable platforms from an early stage. The screen was people who had built early rather than people who had run large mature functions. Those are different jobs, and the second is a poor guide to the first.
Hiring was written into the brief, because the engineers recruited during the engagement would outlast the placement. The commitment was sized to the decisions that needed a senior view rather than to a standard working week, so the business paid for judgment at the points where it changed the outcome.
Architecture set, and a team who could carry it
The platform was built on choices made by somebody who had made them before and knew what each one committed the business to. The more durable outcome was the team: the engineers hired during the engagement understood the architecture because they had been recruited against it.
As the business grew into the role, the arrangement wound down into a permanent appointment that the placement had helped define, rather than being replaced by one. The definition of the job was written by somebody who had already been doing it.
