The Tool That Still Runs Years After I Left

At a gaming platform serving millions of tournament entries, engineer onboarding was a manual checklist: repo access, IAM roles, cloud provisioning, all done by hand, all error-prone, all stealing senior time.

I built an automation tool in Terraform and Go that eliminated 100% of the manual steps. Years later, it is still in production. That is the metric I care about most, more than the observability rollout that cut incident response from 2+ hours to 36 minutes, because it proves the thing consultants rarely prove: the system worked without me.

This is my test for infrastructure work, mine or anyone's. Not 'did it work at launch' but 'does it still run after the author leaves.' If your platform depends on one person's presence, you do not have infrastructure. You have a hostage situation with extra YAML.