What the job actually is

You build and run the compute, network and storage layer that product teams deploy onto — clusters, provisioning, networking, identity, multi-tenancy and the cost of all of it. Unlike SRE this is predominantly build work rather than operate work, and unlike archetype 4 your customer is the system rather than the developer.

Titles this hides behind

Prepare like this

• Know Kubernetes past the manifest. Scheduler behaviour, admission control, what actually happens on a node under memory pressure, and why your pod is pending. Interviewers probe the layer below the one candidates studied.
• Be fluent in one cloud and conversant in a second. All three sit at roughly 50% of postings, so the credible position is depth plus translation, not claimed parity across three.
• Bring a migration story. Something moved, with the rollback plan you prepared and whether you used it.
• Learn the cost dimension — only 5% of postings name FinOps, but it comes up in nearly every senior conversation because it is how platform teams justify themselves.
• Go matters here at 35%, more than anywhere else in either family.

Be ready for

Compensation