Dmitry Yackevich
FLO HEALTH · PLATFORM · 2 MIN READ

A delivery stream from zero

0 → 1
delivery stream, built
Company
Flo Health
Role
Director of Engineering, Infrastructure
When
Oct 2018 – Sep 2022
Result
dedicated delivery stream, 0 → 1

There's a moment in every scale-up where deployment stops being a command and becomes a ceremony. At Flo, releases had quietly turned into that: a ritual with a priesthood — a handful of heroes who knew the incantations, and a queue of engineers waiting behind them.

Heroes don't scale. Worse — they burn out precisely because they're good.

If shipping requires a hero, you don't have a process. You have a single point of failure with a salary.

The fix wasn't a tool. It was an org decision: we stood up a dedicated delivery stream — a team whose product was the path to production itself. Not a "DevOps team" that does deployments for others; a platform team that makes deployments boring for everyone.

They owned the golden path: CI/CD pipelines as a product, environments on demand instead of a shared-staging knife fight, trunk-based development instead of long-lived branch archaeology. Every improvement they shipped multiplied across every product team.

The hardest part wasn't the pipelines — it was the boundaries. Deciding what the platform owns, what product teams own, and where the interface sits. Get that wrong and you've built a ticket queue; get it right and you've built leverage.

Releases went from a scheduled event people planned around to a non-event that happened all day long. Nobody remembers the last "release day" at that point — which is exactly the point.

Next story →
Observability from day one

Working on the same problems? Let's talk.

✉ Email me