Per-MW pricing, regional variance, and cost drivers for owners scoping hyperscale & AI builds.
Salary benchmarks across the 14 mission-critical disciplines.
I’d hire a Cx engineer based on verified test work - not an NFPA 75 keyword. My starting point: two relevant projects, one approved redacted work sample, and two direct references who can confirm what the candidate personally handled.
Here’s how I’d check whether that experience fits your project:
For a live-site retrofit, I’d put extra weight on impairment controls, stop-work decisions, and proof that systems were restored. Knowing the code is only part of the job. <u>The hiring decision should rest on safe testing and traceable results.</u>
Hiring Cx Engineers with NFPA 75 Experience: 4-Step Checklist
Turn the project scope into clear hiring criteria. Start with a scope sheet that spells out the experience candidates need: facility type, operating model, IT load, occupancy, construction phase, criticality tier, and expansion plans.
NFPA 75 protects IT equipment and IT equipment areas from fire, smoke, heat, corrosion, and water. Identify the spaces the engineer will cover, including data halls, electrical rooms, UPS rooms, battery rooms, and support spaces. For phased or modular projects, require experience with factory and site acceptance, module connections, firestopping, and temporary conditions.
Next, define specific interface requirements. Cover fire-resistance-rated barriers, penetrations, firestopping, raised floors, concealed spaces, and enclosure continuity. Require the engineer to map detection, sprinklers, preaction systems, clean-agent systems, fire alarm controls, smoke dampers, HVAC shutdowns, equipment shutdowns, and leak detection to the approved sequence of operations.
Treat UPS rooms, battery areas, and liquid- or immersion-cooling systems as separate scope items. State whether the engineer must coordinate protection measures for ventilation, thermal-runaway response, leak detection, containment, emergency response, and operating procedures.
Require a compliance matrix before test planning. It should list the adopted NFPA 75 edition, applicable codes, manufacturer instructions, issued drawings, permit conditions, owner requirements, insurer requirements, and AHJ interpretations.
NFPA 75 does not replace installation, acceptance-testing, or maintenance standards. Verify governing editions and local amendments. Don't assume the newest edition automatically applies.
Put responsibilities in the job description so you can screen candidates against them. The Cx engineer plans and documents verification, coordinates witnesses, and tracks deficiencies. Designers retain design responsibility, contractors perform acceptance testing, and the AHJ retains approval authority.
Require a clear process for escalating conflicting requirements. Also assign responsibility for impairment planning, emergency response, recovery procedures, and maintenance handoff. Commissioning records do not replace licensed design or AHJ approval.
Use the scope and code map in the next step to check resume claims and verify project experience.
Ask candidates to describe at least two relevant projects, using the same checklist for both. Look for data center fire-protection work on hyperscale, colocation, enterprise, or modular projects. Record dates, project phase, systems tested, and the deliverables each candidate owned. Separate work they authored, executed, managed, witnessed, or reviewed. Attending a test does not prove ownership.
Request one redacted work sample covering pre-functional checks, functional performance testing, or integrated systems testing. Trace the sample through submittal reviews, cause-and-effect matrices, test records, inspection reports, and deficiency closure. Verify experience with NFPA 72, NFPA 13, NFPA 2001, and NFPA 4. NFPA 4 covers integrated systems testing.[6][7] Treat NICET certification, commissioning credentials, inspection qualifications, and PE licensure as supporting evidence - not replacements for documented work.
Use the table below to tie each claim to a traceable project record.
Have candidates fill in the applicable rows with actual project details. Allow anonymized identifiers at first. Each project entry should include facility type, square footage, redundancy configuration, dates, and phase. Each responsibility entry should separate procedure authorship, execution management, witnessing, and record review.
Use these entries to ask how the candidate handled actual test windows and deficiencies.
Ask candidates to walk through one actual test window involving the owner, operations, general contractor, fire-protection and fire-alarm vendors, controls, and electrical and mechanical trades. Get specifics: which prerequisites they checked, which participants they confirmed, how they coordinated impairments, and how they assigned responsibilities. Then trace one deficiency from discovery through correction, retesting, and the turnover recommendation.
Choose an incident involving an unavailable vendor, incomplete installation, missing acceptance records, or a failed test. Look for judgment under live-site constraints - not attendance. Ask what stopped, what changed, and what evidence supported retesting. Confirm who received the escalation and how the team controlled life-safety and business-continuity risks. Separate missing paperwork from a functional failure; neither should be marked closed without acceptable evidence.
After checking project history, test whether the candidate can apply code requirements to live commissioning decisions. Knowing NFPA 75 is not the same as being able to commission a system.
Use a panel with commissioning, fire-protection, controls, construction, and operations experience. Score candidates on prerequisites, sequences, acceptance criteria, safety controls, and records - not their ability to recall standard names.
Start with the highest-risk interfaces: detection, preaction, clean-agent release, and HVAC shutdown, including restoration. Ask candidates how they would verify ceiling-level and underfloor detection while accounting for airflow, obstructions, detector locations, and aspirating-system sampling points. They should clearly separate visual checks, device activation, and full interface testing.[5][8][10]
For clean-agent release logic, require a safe test mode and impairment controls approved by the vendor and owner. Candidates should verify coincidence logic, delays, abort/manual release, notifications, HVAC or damper response, and documented restoration.[11][4] A simulated release does not prove suppression acceptance.
Sequence testing and acceptance testing need separate evidence. Depending on the system, candidates should identify sprinkler piping and flow-test records, preaction trip results, or clean-agent quantity and enclosure-integrity reports. Each interface check must link to the cause-and-effect matrix. Each suppression record must meet its own acceptance criteria.[4][8][9]
Use three scenarios:
An unresolved fire-protection fault before integrated testing. A UPS/battery area requiring detection, ventilation, and shutdown verification. A modular unit whose risk assessment differs from the baseline design.
Look for candidates who identify unsafe test conditions, check battery chemistry or module protection boundaries, and work through the designer to reconcile adopted codes, approved documents, and AHJ direction. They must not redesign systems or waive requirements.
Then ask them to walk through one failed test: the expected result, observed failure, party responsible for correction, and dated retest evidence. A claim that “the vendor resolved it” is not enough without supporting records.
Finish by asking candidates to turn their judgment into a workable test plan. Give everyone the same occupied or near-turnover facility scenario. Require a short plan covering readiness, sequences, pass/fail criteria, instruments, vendor assignments, operations approval, impairment controls, stop-work conditions, restoration, and turnover records.
Score four categories from 0–5 each: code/document interpretation, safe planning, cross-trade coordination, and objective records. Reserve scores of 4–5 for traceable answers that name the governing matrix, assign each check, define results, and prove normal system status after testing.[12] Look for evidence that the engineer can run a safe test window and document the work.
Use the test-plan scorecard to check whether the candidate can close out projects - not just discuss code. Request redacted work samples approved for sharing from comparable data center projects. These should include test scripts, functional performance test forms, cause-and-effect matrices, deficiency logs, integrated systems testing records, and closeout reports.
Ask candidates to identify what they authored, executed, reviewed, or witnessed. Look for records that connect test criteria to observed results, corrective work, retest evidence, and closure approval. These records should show that the candidate can carry out commissioning work in a live data center without creating uncontrolled risk.
With consent, contact two references who directly supervised the candidate. Give priority to references who can confirm that the candidate stopped unsafe testing when conditions changed. Check authorship, vendor coordination, and turnover milestones against project dates and submitted records.
Once references confirm ownership, score the candidate by system area. Record verified strengths and gaps across detection/notification, water-based suppression, preaction, clean-agent or special-hazard systems, fire pumps, monitoring interfaces, emergency power, HVAC shutdown, dampers, access control, and BMS integration. Score gaps only for systems the project uses.
Confirm familiarity with the project’s adopted NFPA 75 edition, AHJ process, permit conditions, owner requirements, and site operating constraints. Verify required licenses, manufacturer training, and credentials. Use NICET as supporting evidence, not the primary basis for judging qualifications.
Document the facility type, phase, location, schedule, shift and travel expectations, required systems, deliverables, vendor interfaces, and approval authority. Match the brief to the project scope rather than using a generic commissioning job description.
Separate mandatory capabilities from trainable gaps. Treat software gaps as trainable and required-license gaps as disqualifying. Attach the assessment scorecard, work-sample requirements, reference findings, and a mitigation plan for accepted gaps. Use the final brief to get hiring managers aligned before offers go out.
iRecruit.co offers commissioning recruiting, RPO, and consulting support for mission-critical construction teams.
PE licensure isn’t required for every commissioning task. The need depends on the role’s responsibilities and risk profile. A license can strengthen a candidate’s profile and is often linked to senior roles [1][2][3].
But recent mission-critical field experience typically carries more weight than licensure alone. Look for documented startups, integrated systems testing, and deficiency tracking. Focus on verified project results and hands-on commissioning expertise [1][2][3].
Use scenario-based interviews to assess technical depth and decision-making under pressure. Ask candidates to walk through failures they managed, including their role in root-cause analysis, escalation, corrective actions, and retest criteria. Listen for clear ownership:
“I led” or “I wrote.”
Request redacted test scripts, deficiency logs, or handover documents to review their work. Then use professional references to confirm who wrote the procedures, managed the issue log, and handled handover responsibilities.
New hires need verified expertise in the electrical, mechanical, and controls systems they’ll test, along with documented L1–L5 commissioning experience. They must also know how to read one-line diagrams, control schematics, and sequences of operations.
The role includes managing issue logs, documenting defects, and closing out retests. Safety is pass/fail: hires must have documented OSHA 30 training, plus NFPA 70E training for work around energized equipment. They must also understand lockout/tagout and arc-flash boundaries.
