Loading…
Also forInfrastructure EngineerDevOps EngineerPlatform EngineerCloud Infrastructure EngineerDeveloper Platform EngineerCloud engineer loops — whether the title on the job ad says cloud, infrastructure or platform engineer — assess whether you can build and run the layer other engineers work on top of, which is a different judgement from building a product. Expect real detail on networking and identity, on how infrastructure is defined and applied and what happens when someone changes something by hand, and on delivery: pipelines, rollouts, rollback. The design questions are less about product features and more about blast radius, guardrails, multi-tenancy and the order of operations during a change; the recurring theme is that a platform nobody can work within will be routed around, so answers that only optimise for control score badly.
A design question on account structure, network topology, a migration or a delivery pipeline; strong emphasis on infrastructure as code including state, drift and safe apply; deployment strategies and rollback; orchestration and compute model choice; guardrails and multi-tenancy; observability as a platform capability rather than a per-service concern; and at least one troubleshooting scenario, usually spanning more than one system, where you are handed symptoms and expected to narrow them methodically. Security comes up throughout rather than as its own round.
Two, and they are the same mistake from opposite ends. Describing infrastructure as code without ever mentioning state, locking or drift — the first thing an experienced interviewer probes, because it is where the operational difficulty lives. And designing guardrails without considering how teams onboard: excellent controls plus a two-week path to an environment produces shadow infrastructure with no controls at all, and interviewers with platform experience listen for whether you know that.
Prefer a sequence to a catalogue?
The 6-week Cloud Engineer study plan orders this material week by week, easiest first, with a free first week and progress derived from what you have already completed.
Fully unlocked, solutions included. Enough to see whether the format suits you.
Broad questions with no system to draw. Free, and frequently where this loop is decided.
Easiest first. 48 in total for this role.
Symptoms with no cause attached. 6 in total for this role, easiest first.
At senior level it eliminates more candidates than the technical rounds do. Free interactive guides to the questions that come up most.
Close enough to prepare for together, which is why this is one path. All three loops centre on infrastructure as code, networking and identity, delivery and rollback, and a troubleshooting scenario. Platform-titled loops lean a little further toward developer experience and the internal services other teams build on; infrastructure-titled loops a little further toward the network and identity layer. The questions here cover both ends.
Typically scripting and automation rather than algorithm puzzles — parsing output, automating a repetitive task, or reasoning about a configuration. Some loops include a standard coding round; it is rarely the round that decides the outcome.
Very often, and it is worth preparing for specifically. You are given a symptom and expected to narrow it methodically rather than to guess a cause. Establishing what the words mean before theorising is what earns credit.
They overlap heavily and many loops are near-identical. Cloud and platform loops lean toward building internal capabilities and the infrastructure layer; reliability loops lean toward operating production, incident response and the measurement of reliability itself. The SRE path weights capacity arithmetic and SLOs where this one weights infrastructure as code and delivery.