Hiring a Kubernetes Engineer?
Your team is hiring; I run the technical round for you. One 60-minute live interview per candidate, with a scored written recommendation back within 12 hours.
About the role — Runs clusters in production, not in a tutorial.
What you get
A 60-minute live interview with a scored written recommendation back within 12 hours. No retainer; fee quoted on the scoping call.
01 — Why this role is hard to assess
Nearly every candidate has run a cluster. Far fewer have debugged one under load, upgraded one without downtime, or explained to a finance team why the node group costs what it does.
02 — What the 60 minutes covers
The specific ground, not a generic checklist.
- A cluster upgrade they ran on live traffic, and what went wrong
- Resource requests and limits: how they set them and how they know
- Debugging a CrashLoopBackOff with no useful logs
- Autoscaler behaviour under burst — Karpenter or Cluster Autoscaler, and why
03 — A scenario from the round
Pods are being evicted on one node group during peak hours. Requests and limits look fine, the nodes are not obviously saturated. Take me through it.
Scenarios rotate between candidates so the round stays fair, and none of the published examples are used verbatim.
What I listen for
Failure modes specific to this role
- Has only ever used a managed cluster someone else configured
- Sets limits equal to requests without being able to say why
- Reaches for more nodes before understanding the eviction
- Cannot explain what the scheduler actually did
Scored on
- Cluster operations
- Scheduling and resource reasoning
- Debugging under pressure
- Upgrade discipline
- Cost awareness
Each dimension is scored with written rationale. You get the reasoning, not just the number.
Next step
Hiring a Kubernetes Engineer this quarter?
Fifteen minutes to agree the rubric, then I run the round. One candidate or twenty — there is no retainer and no minimum.