Infrastructure for several startups, built from zero.

The problem

Startups need production-grade infrastructure long before they can afford a platform team. Running several of them at once means every setup has to be quick to create, safe to run, and cheap enough for a company that hasn’t found its customers yet.

What I changed

I set up each startup’s infrastructure from scratch, with CI/CD and the tooling around it, and managed them side by side. Doing it more than once is the point: patterns that worked became the starting point for the next one.

The result

Each startup had production-ready infrastructure from its first release, with cloud costs optimised to its stage instead of sized for a scale it didn’t have yet.

Step by step

1. A repeatable starting point

   Each new startup began from the same proven setup instead of a blank page, so day one looked like day one hundred.

2. CI/CD from the first commit

   Every product shipped through a pipeline from its first release, with the tools around it already in place.

3. Cloud bills that match the stage

   Costs kept in line with each startup's size and traffic, and trimmed as usage became clear.