DevOps engineer, SRE or platform engineer. Which one are you hiring?
Three job titles that get used as if they were one. They are not, and choosing the wrong one costs you about a quarter.
Most of the briefs we receive open with a job title and close with a description of a different job. The title says DevOps Engineer. The description says someone to define service level objectives, run incident reviews and reduce pages at night. That is an SRE. Advertise it as DevOps and you will interview a lot of capable people who want to build pipelines.
The titles overlap, and plenty of engineers have held all three. But each one describes a different centre of gravity, and good candidates choose jobs by it.
DevOps engineer
Owns the path from a commit to production. Pipelines, infrastructure as code, environments, the deployment itself. The strong ones are generalists by temperament. They enjoy being the person who knows how everything fits together, and they measure themselves by how rarely developers need to ask them for anything.
You probably need one if releases are manual, slow, or depend on a single person who is not allowed to take holiday.
Site reliability engineer
Owns what happens after the deploy. Availability, latency, alerting, on-call, incident response and the review afterwards. An SRE treats reliability as an engineering problem with numbers attached, and expects to write real software to solve it.
You probably need one if your product is already in production at some scale, outages now have commercial consequences, and nobody can say what your availability was last month without opening a spreadsheet.
Platform engineer
Owns the internal product that other engineers ship on. Usually that means Kubernetes and everything around it, wrapped in enough tooling that a product team can get an environment, a pipeline and sensible defaults without filing a ticket.
You probably need one, or more likely several, once you have enough product teams that each of them is solving the same infrastructure problems separately and slightly differently.
How to tell which one you need
Ask what would be different in six months if the hire went well.
If the answer is “we ship more often and it is less frightening”, that is DevOps.
If the answer is “we have fewer incidents, shorter ones, and we stopped being surprised by them”, that is SRE.
If the answer is “our developers stopped waiting for us”, that is platform.
If the honest answer is all three, you have two options. Hire a senior generalist and accept that they will do one of the three properly for the first year. Or admit you are building a team, and decide which hire comes first. For most companies below a hundred engineers, the order is DevOps, then platform, then SRE. Reliability specialists are most useful once there is a stable platform to make reliable.
Why the title matters to candidates
Senior infrastructure engineers read job titles as a statement of what they will be allowed to do. An SRE who joins a team expecting to own reliability and spends the year maintaining Jenkins will leave, and will be right to.
Get the title and the first paragraph honest and you will receive fewer applications. They will be from people who want the job you actually have.