CI
What should a good CI pipeline optimise for?
Whether the candidate treats CI as a feedback system with a latency budget, or as a list of tasks that must all be run.
The situation behind the question
A service has one CI job: install dependencies, run the full test suite, build a Docker image, run end-to-end tests against a shared staging database, then push the image. It takes 38 minutes, fails roughly a third of the time for reasons unrelated to the change, and the team has a shared habit of hitting re-run twice before reading the log.