Hiring guide · CNC machinists
Machinist skills test: what to measure and how to score it
A machinist skills test should measure the handful of skills that decide success in the actual opening — typically setup and prove-out, program edits at the control, print reading and GD&T, in-process inspection, and troubleshooting to root cause — with each answer scored against documented criteria defined before testing begins, so every applicant is measured identically.
Measure judgment, not vocabulary
Anyone can memorize what SFM stands for. What separates machinists is judgment under real conditions: what to check before cutting a rework part that welding may have pulled, when a correction belongs in a wear offset versus the program, what a shallow pitch diameter says about a threading cycle. Questions framed as shop scenarios surface that judgment; definition quizzes do not.
The scoring criteria are the test
Before any applicant answers, write down what a strong answer must contain. For a threading-cycle question that might be: corrects depth from measurement rather than guesswork, considers the wear offset for a fine correction, restarts mid-program safely, re-measures before running the batch. Scoring then becomes checking each answer against those criteria — the same way for everyone.
This is what makes results usable. 'I liked the second guy' is not comparable across applicants and not defensible later; 'demonstrated four of five criteria, did not demonstrate re-measurement' is both.
What a scored report should show
Per applicant: which skills were demonstrated, which were partial, which were not demonstrated — mapped to the requirements the employer named, nothing decorative. Across applicants: a comparable rating, so the hiring decision starts from evidence. Every finding should trace to a specific answer, which keeps the process auditable.
Common questions
Can I write a machinist test myself?
Shop owners do, and a self-built test beats no test. The common gaps are scoring drift (each applicant judged a little differently), questions that leak the answer, and no documentation of job-relatedness. If you build your own, define written scoring criteria per question before the first applicant takes it.
Should the test be different for every opening?
Yes. A 5-axis prove-out question is irrelevant to a manual-lathe opening. Building the assessment from the specific job posting keeps every question job-related — which is also what selection guidelines expect.
How is HireSciences' scoring kept objective?
Scoring is deterministic: documented rule-based criteria per question, applied identically to every applicant, with each report reviewed for completeness and consistency before release. Findings map only to the requirements the employer names at intake.
See it on your own opening
See how this works for machine shops, try the 2-minute demo, or take one of the two paths below.