At some point every startup asks it: is it time to hire a DevOps person? Having sat on both sides of this, my honest answer is that founders usually ask the question about a year before a full-time hire makes sense, and then, once it does make sense, take a year to act on it. Here's how I'd think it through.
The real signals it's time
- There is consistently 40 hours a week of platform work, not 40 hours of backlog you wrote down once, but recurring weekly demand. Be honest; most startups under 25 engineers don't have it.
- Product engineers are losing a day or more a week to infrastructure toil, and it's getting worse, not better.
- You're heading somewhere structurally demanding: multi-region, heavy compliance, on-call that actually pages.
The expensive ways to get it wrong
Hiring too early is the classic. A good platform engineer costs $180k+ loaded in Australia, and if there's only fifteen hours of real work a week, one of two things happens: they get bored and leave, or they generate work to fill the time, and you end up with a Kubernetes cluster for a product that three EC2 instances were serving fine. Both outcomes cost more than the salary.
The second classic is hiring junior to save money. Your platform is the thing that keeps customer data safe and the product online. It's the last place in the company to learn on the job unsupervised.
The middle options nobody mentions
Between "founder does everything at midnight" and "$180k hire" there's a wide, sensible middle. Fractional senior help for the recurring ten hours that actually exist. Fixed projects for the lumpy work like migrations and compliance. Levelling up a keen product engineer with senior guidance behind them, which is often the best long-term answer, because in eighteen months they become your first platform hire, from the inside.
If you do hire
Look for breadth over speciality; a startup platform person touches networking, CI, security and cost in the same week. Prefer a bias for boring technology, and treat "writes things down" as a first-class skill, not a nice-to-have. And get someone senior into the interview loop, because platform engineering interviews are very easy to bluff if nobody in the room has done the job.
Not sure which side of the line you're on? That's an advisory hour, and I'll happily tell you if the answer is "don't hire anyone, including me". Ask the question.