Sébastien, co-founder of Lunnar.

Seven years keeping some thirty companies running

I spent seven years as a systems and network administrator, on Linux and on Windows, inside a group that brought together more than thirty companies. Thirty different trades, so thirty ways of working, thirty vocabularies, and as many constraints with nothing in common.

It is a schooling you do not get inside a single company. When you hold up the infrastructure of thirty businesses at once, you learn fast what separates a tool that holds from one that buckles at the first load spike. You also see what it costs, in hours and in frustration, to run a system built for a generic trade and bent to fit a real one.

Projects run end to end, not just the technical part

Across those years I did more than keep servers alive. I ran projects: drawing the need out of people who have no technical vocabulary, deciding what gets built and what does not, following the studies, holding deadlines, building websites and keeping them alive after launch.

That is where I learned the point that comes back in nearly every article I write: a project scoped badly at the start is never rescued by coding faster. It gets paid for later, always at a higher price, and it is the client who pays.

What that changes about what you read here

When I write that a quote has to spell out the integrations, the level of security and what happens after go-live, that is not a theoretical position. It is the list of points where I have watched projects go sideways, from the side of the person who then had to keep the whole thing running.

And when I write that security is not something you add at the end, it comes from the same place: I have repaired it after the fact often enough to know what that costs, next to laying it into the foundations.

Today

I am a co-founder of Lunnar alongside Guillaume. We both work across the whole of a project, from design through to production. At two people you cannot afford sealed borders, and in any case a project where each of us only saw half of it would be badly designed: architecture decisions and usage decisions get made together or not at all.

If my name is attached to architecture and security, it is because that is where I am most useful, not because it is fenced off. Guillaume has the same relationship with design and construction.

There are two of us, and we say so on purpose. You know who is working on your project, and there is nobody standing between you and the people building it.