What CNC Machinist Interviews Actually Test

Who's actually asking the questions

For most machine shops, the person across the table is not HR. It's the shop supervisor, the lead machinist on your shift, or the shop owner if it's a smaller operation. In aerospace or medical shops running to AS9100 or ISO 13485, a quality inspector or quality manager often sits in as well, because their sign-off matters as much as the supervisor's. HR, where it exists at all, usually just handles scheduling and pay talk. This matters because the person deciding whether you get hired has stood at a Haas VF-2 or a Mazak with a bad tool and a part on the line, and they are listening for whether you have too.

That changes what "good" sounds like. A generic answer about being a "team player" or "detail-oriented" gets you nothing here. The supervisor already knows those words. What they don't know yet is whether you can indicate in a vise inside two tenths, whether you'd catch a wrong work offset before it ruins a $400 forging, or whether you know the difference between a problem you fix yourself and one you flag before it becomes six scrapped parts.

The practical assessment

Most shops that are serious about hiring will run some form of hands-on test, and it usually falls into one of two shapes.

A machine test. You're given a print, a blank, and access to a mill or lathe — often the same brand and control you'd be running day to day (Fanuc, Haas, Siemens, Okuma, Mazatrol). You might be asked to set up a vise and tram it, pick up work coordinates in G54, touch off tools, and either run an existing program or write a short one to cut a feature to print. Some shops hand you a program with a deliberate error in it — a wrong tool offset, a feed rate that would crash on the first plunge — and watch whether you catch it before you hit cycle start. This is not really about whether you can cut the part. It's about your sequence: do you dry-run it, do you check the tool list against what's actually in the carousel, do you run the first piece in single block with the feed override backed off. Someone who's actually done this work will do these things without being told to.

A print and math test. No machine, just paper. You'll be asked to read a blueprint with GD&T callouts — true position, flatness, perpendicularity — and explain what they mean for how you'd hold and machine the part. You might be asked to calculate speeds and feeds for a given material and tool diameter, or to spot an error in a setup sheet. This tends to show up more at shops running tighter tolerances or higher-mix work, where reading a print correctly the first time is the job.

Either way, the test is rarely pass/fail on the finished part alone. A shop that's paying attention is watching your habits: do you measure with the right instrument for the tolerance (a caliper is not a micrometer, and neither is a comparator or CMM), do you deburr and inspect before you walk away, do you log what you did.

The questions that are really testing something

A few questions come up again and again, and they're not small talk. Here's what they're actually probing.

"Walk me through how you'd set up this job from a cold start." They want your sequence, not a summary. Blank stock check, fixture or vise selection, indicating it in, picking up work zero, tool list against the program, first article inspection, then production run. If you skip straight to "I'd load the program and hit go," that's a red flag to anyone who's had a part run into a chuck.

"What do you do when a tool breaks mid-cycle?" This is testing judgement under a real, common failure, not troubleshooting trivia. A strong answer covers: stop the spindle, don't just hit reset, check the part and the remaining tools for damage, determine whether the part is scrap or reworkable, re-touch off the replacement tool, and document it — not just fix it and move on silently. A shallow answer is "I'd just put a new tool in and carry on," which skips the part where you find out if the crash already ruined the workpiece or chipped the next tool in the carousel.

"How do you know a program is safe to run at full speed the first time?" They're checking whether you understand dry runs, single block, and feed override — not whether you're brave. Someone who says "I'd just run it" either hasn't been burned yet or is telling you what they think you want to hear.

"Tell me about a time you caught a bad part before it shipped." This is about your inspection habits under normal production pressure, not a heroic story. Good answers name the actual check — a go/no-go gauge, an SPC chart trending out of control, a first article that didn't match the CMM report — and what you did about it, including telling someone even when it slowed the line down.

"What controls have you run, and what's different about programming on them?" If the job is on a Fanoc control and you've only run Mazatrol conversational, that's fine to say plainly. What isn't fine is implying G-code is G-code everywhere. Canned cycles, macro variables, and how offsets are entered differ enough between Fanuc, Haas, and Siemens controls that anyone who's actually run more than one knows it.

What a shallow answer sounds like

To someone who has run these machines for years, certain phrases are an immediate tell.

"I always double-check my work" with no instrument named. Which instrument, at what tolerance, at what interval? A machinist checking a ±.0002" bore with a dial caliper hasn't checked it at all.

Talking about "programming CNC machines" without naming a control. Every shop runs specific hardware and software, and if you can't say Fanuc 0i, Haas NGC, or Mazatrol, it reads as if you've watched it done rather than done it.

Not knowing what G54 through G59 are for, or claiming you "never really needed offsets" — that only makes sense if every job you've run used a single fixed fixture, which is unusual outside very narrow production work.

Describing scrap or a crash with no mention of what you changed afterwards. Everyone crashes a machine eventually. What the interviewer is listening for is whether you learned the specific cause — wrong tool length offset entered, program run in the wrong work coordinate, fixture not fully seated — not a vague "I learned to be more careful."

What to actually do before you go in

Bring your own tools if you have them — calipers, mics, an edge finder, a dial indicator — even if the shop says they'll provide equipment. Using your own gear is a small but real signal that you work this job outside of just this interview.

If you hold a NIMS credential, a specific machine certification, or documented CNC apprenticeship hours, have the certificate number or issuing body ready, not just the title.

Before the interview, pull up the specific control the job posting or the shop's equipment list mentions and refresh yourself on its quirks — how it handles tool wear compensation, how canned cycles are called, what its alarm codes look like. If you're rusty on GD&T symbols, spend an evening on it; a shop running tight tolerances will hand you a print with true position and flatness callouts and expect you to talk through them without hesitating.

And if you genuinely haven't run the exact control or material they're asking about, say that plainly and say what you have run instead. A machinist who's honest about the boundary of their experience reads as far more hireable than one who bluffs it and gets caught on the shop floor.