Home/Resources/How to technically screen a network engineer

Resources · Hiring and architecture

A candidate who reads well and interviews well can still be unable to reason under pressure.

You usually discover a bad specialist hire in month three, on an incident. These are the questions that move that discovery to before the offer — usable whether or not you work with us.

Why resume screening cannot do this

A recruiter cannot assess whether a candidate understands a dial-plan conflict, and nobody expects them to. Keyword matching is the only screening available at that stage, so the technical filtering necessarily happens later — in your interview room, performed by your senior engineers.

That is the actual cost of a resume-first process. Not a bad hire in itself, but your most expensive people spending the week interviewing instead of engineering, and the requisition staying open while the project it was raised for slips.

What is worth assessing

Four dimensions, and the order matters. Depth without judgment produces someone who knows the platform and cannot be trusted with a change window.

  • Hands-on depth — has this person operated the environment they are being submitted for, in production, rather than studied it?
  • Protocol fundamentals — can they reason from first principles when the symptom does not match the documentation?
  • Failure-handling judgment — what they do when a change goes wrong, and whether they have a rollback instinct or a retry instinct.
  • The ability to explain a decision — because an engineer who cannot articulate why will not be trusted by the team they join.

Questions that separate candidates

The useful questions are not trivia. They ask the candidate to reason about something going wrong, because that is the condition under which the hire actually matters.

  • Describe a change you made that went wrong. Listen for whether they know what they would check first next time, or whether the story has no diagnosis in it.
  • For an automation-adjacent role: what would you refuse to automate in a live network, and why? A candidate with no answer has not worked at production scale.
  • If a change is pushed to 200 devices and the twelfth fails, what happens to the other 188? Good answers involve state validation and rollback, not retries.
  • Walk me through diagnosing one-way audio. This separates people who have handled voice faults from people who have read about them.
  • What in your current environment is undocumented, and how do you work around it? Honest answers here predict how they will handle yours.

Scope the role before you screen for it

Half of failed specialist hires are scoping failures rather than screening failures. These four questions usually reshape the requirement before anyone is interviewed.

  • Is the specification describing the person who left, or the person you need next?
  • Which requirements are genuine and which were added because someone thought they should be there? Over-specification is the most common reason a good hire is filtered out.
  • Is this a capability gap or a capacity gap? They look identical on a requisition and require different people.
  • What is the actual seniority of the work — an architect's problem being hired at engineer level, or the reverse?

The third question is the one that changes engagement shape rather than just the job description. A capacity gap with a defined end date and a permanent capability gap are different instruments, and hiring permanently for the former is how organizations end up carrying headcount they did not need.

Related

Where to go next.

Scope the role with someone who could do the job.

Thirty minutes to define the requirement properly, then which track fits and what it costs.