Most bad mission-critical hires interview well. Titles on a resume tell you where someone stood, not what they were responsible for, and 'data center engineer' spans a strong electrical technician through to a site engineering lead. The framework below is the sequence we use before a candidate reaches a client shortlist.
1. Establish the environment, not the title
Ask what the site was: hyperscale, colocation, or enterprise; how many halls; what the load was; whether the space was live while they worked in it. A candidate who has only worked single-tenant enterprise will find multi-tenant colocation, with customer-facing work and cross-connects, a genuinely different job.
2. Verify the plant, system by system
Name the equipment and listen for specifics. Which UPS platform, which switchgear, which generator and ATS arrangement, which building management system. On the mechanical side: chilled water or DX, CRAH or CRAC, air-side configuration, and any liquid cooling exposure. Vendor-specific familiarity with the plant you actually run shortens the time before a new hire can stand watch alone.
3. Test procedure discipline with a scenario
Mission-critical work is procedural. Ask the candidate to walk through a real method of procedure they wrote or executed, then give them a scenario: an alarm on a UPS during a planned maintenance window, with a customer on the floor. You are listening for change control, escalation, and the instinct to stop work — not for a hero story about fixing it alone.
The strongest answers include a time the candidate halted a procedure. People who have carried the responsibility have all done it. People who have watched from nearby rarely mention it.
4. Confirm the practical reality before the offer
Shift rotation, on-call frequency, physical requirements, travel between sites, and background or clearance requirements all need confirming in the process rather than at offer stage. Certifications and licences should be verified rather than accepted from the resume. These checks cost minutes and prevent the late decline that restarts the whole search.
5. Red flags worth pausing on
Vague answers about plant they claim to have operated. Short tenures with no stated reason across several live sites. Discomfort discussing procedure and change control. An unwillingness to name a mistake. None of these disqualify on their own; all of them warrant a second conversation before an offer goes out.
If you would rather this screening happened before resumes reach your inbox, that is the work we do. Send us the role and we will come back with a shortlist that has already been through it.
Hiring for a data center role?
Tell us what you are hiring for and we will come back with availability, comp ranges, and next steps. No fee to talk.