DevOps Engineer interview loops

A DevOps engineer is hired so that changing production stops being an event. Other engineers merge a pull request; what happens between that merge and customers running the change is this person's work — the pipeline that builds and tests it, the deploy that puts it on servers, the infrastructure code those servers come from, and the credentials each of those runs with. The title covers very different jobs at different companies, so the loop here takes the part they share: the path from a commit to production. Service levels, alerting and on-call are a site reliability engineer's to interview for, and the layout of a cloud estate across accounts is a cloud architect's.

The trouble with interviewing for this role is that its vocabulary is easy to acquire. Immutable artifacts, blue-green, least privilege and drift detection can all be recited in the right order by someone who has never owned a pipeline. What cannot be recited is the specific defect in the thing on the table: a pipe that turns a failed build into a green deploy, sixteen planned changes of which one deletes the production database, a canary gate that promotes a release because the release is the only thing still working. So every round below gives the candidate an artifact with defects the interviewer planted and knows, and scores what the candidate finds and what they do about it.

A loop is a set of round plans in order. Each round scores one to three competencies, every must-have is scored in at least two rounds, and the debrief is held competency by competency rather than round by round, so no single interview decides the hire. How these plans work covers the rest of the method.