60 Translator interview questions to ask job applicants

Ask a translator to describe a project that went wrong, then ask what they changed afterwards. That single pair of questions separates a careful linguist from someone who edits machine output quickly, and it does more work than any question about language pairs or software. The rest of this page is 60 questions built on that idea, grouped by what each one is meant to prove.

First, a correction to something this page used to repeat. Translator demand is not surging. The Bureau of Labor Statistics projects employment of interpreters and translators to grow 2 percent from 2025 to 2035, slower than the 3 percent average across all occupations, on a base of about 73,900 jobs. Roughly 6,000 openings a year are expected, and most come from people leaving the occupation rather than new headcount. So you are not fighting a talent shortage. You're sorting a steady pool where the quality spread is enormous. That is the job whether you're a marketing agency localizing campaigns, a non-profit publishing in six languages, or a small team hiring your first in-house linguist.

TL;DR

• Machine output is fluent enough that a plain translation test no longer tells you who can actually translate. Design the test so tool use becomes visible instead of banned.
• Interview for judgment, not vocabulary. The useful questions are about ambiguity, client disagreement, confidentiality and what the candidate does when the source text is bad.
• Tie every competency to a piece of evidence you can collect. If you can't name the artifact, the question is decoration.
• Score with a rubric written before you read the first submission. Research on selection consistently favors structured scoring over a panel re-weighing the same evidence by feel.
• For rare language pairs there is no certification and often no second reviewer. Hire on process evidence and a paid pilot, not on a test nobody can mark.

Which translator interview questions actually work?

The questions that work ask for a decision and its consequence, not a capability. "How do you handle idioms" invites a rehearsed answer. "Tell us about an idiom you got wrong, how you found out, and what you do differently now" cannot be answered without a real project behind it. Every question below is built to that shape.

Before you write the list, decide what each competency looks like as evidence. This table is the version worth stealing.

Pro tip: send the passage 24 hours before the interview and spend the call on the reasoning instead of the output. You learn more from ten minutes of "why did you choose that word" than from an hour of general questions.

What general questions should you ask a translator?

These 15 open the conversation and establish scope: which directions the candidate works in, what they specialize in, and what they refuse. Direction matters more than people expect. Plenty of strong linguists work confidently into their native language and shakily out of it.

1. Which language pairs do you work in, and in which direction do you translate professionally?
2. Which subject areas do you take on, and which do you turn down?
3. Do you have work or academic experience in localization, translation, proofreading, copyediting, linguistic quality assurance or similar fields? If yes, please provide details.
4. What was the last document you translated, and who was it for?
5. How many words a day can you produce at the quality level you'd want reviewed?
6. Which translation memory or terminology tools do you work in, and which do you dislike?
7. How do you build a glossary for a client who has never given you one?
8. What do you do when the source text is badly written?
9. How do you handle a term with no equivalent in the target language?
10. Describe your self-review process between finishing a draft and sending it.
11. What does a good brief from a client contain?
12. How do you price and schedule a rush job without damaging quality?
13. Which style guides do you work to, and how do you handle conflicts between them?
14. What confidentiality practices do you follow with client files?
15. What would make you decline a project?

Which behavioral questions reveal real projects?

Behavioral questions earn their place when they force a specific past event. Ask for the project, the constraint and the outcome. A candidate who has done the work answers in detail; one who hasn't answers in principles.

1. Tell us about a translation you'd do differently now. What changed your mind?
2. Describe a time a client disagreed with your word choice. How did it end?
3. Tell us about a deadline you were going to miss. What did you do?
4. Describe a project where the source text contained an error. How did you handle it?
5. Tell us about a time you had to translate something you disagreed with.
6. Describe the hardest terminology decision you've made and how you documented it.
7. Tell us about working with a reviewer whose corrections you thought were wrong.
8. Describe a time you had to ask a client an uncomfortable clarifying question.
9. Tell us about a file that arrived in a format you couldn't work with.
10. Describe a project where cultural context changed the translation, not just the words.
11. Tell us about a time you caught a confidentiality risk in a document.
12. Describe how you handled a client who kept changing the source text mid-project.
13. Tell us about the longest project you've run and how you kept terminology consistent.
14. Describe a time you turned down work you could technically have done.
15. Tell us about feedback that genuinely improved your work.
16. Describe a time you worked with a subject-matter expert to get a passage right.
17. Tell us about a project where the deadline forced a quality tradeoff. What did you sacrifice?

Personality and work-style questions

Translation is solitary, deadline-driven and mostly unsupervised. These 13 questions probe the working habits that decide whether someone is reliable over months, not whether they're pleasant in an interview.

1. How do you structure a working day when nobody is checking on you?
2. What does your workload look like when you're at capacity?
3. How do you decide whether to accept a project that would stretch you?
4. What part of the work do you enjoy least, and how do you get through it?
5. How do you keep your target language current when you live outside the country?
6. What do you read to stay sharp in your specialist field?
7. How do you respond when a client gives vague feedback like "it doesn't sound right"?
8. How do you handle several clients wanting the same slot?
9. What does "good enough" mean to you on a low-budget project?
10. How do you decide when to stop revising?
11. Describe how you prefer to receive corrections.
12. How do you handle a long project that has become boring?
13. What would your last client say was your weakness?

Should translators use AI in a hiring test?

Yes, with a stated policy and a task designed to make the use visible. Banning tools in a take-home test is unenforceable, and it selects for candidates willing to say what you want to hear. The better move is to hand over a machine-translated draft and ask the candidate to post-edit it, then judge the edits.

That reframes the whole assessment. A plain translation task now mostly measures how fast someone can tidy up a machine draft. A post-editing task measures whether someone can spot a fluent sentence that says the wrong thing, which is exactly the failure mode machine output produces and exactly what you are paying a human for. Write the policy into the brief in one line: which tools are allowed, what has to be disclosed, and that the interview will cover their edits.

1. Which machine translation or AI tools do you use, and at which stage of your process?
2. Here is a machine-translated passage. Post-edit it, and mark every change you'd defend to a client.
3. Which errors does machine output make in your language pair that a non-speaker would never notice?
4. When would you refuse to post-edit and insist on translating from scratch?
5. How do you handle a client who asks for machine quality at machine prices?
6. What do you do when a client's terminology database conflicts with what the tool suggests?
7. How would you explain to a client why your rate is higher than a machine's?

The Testlify Assessment Integrity Framework is the piece that makes this practical. It protects the trustworthiness of results through identity checks, proctoring controls, AI assistance detection, suspicious-behavior signals and reviewable evidence, while the final judgment stays with a human. Applied to a translation test, it means you get the submission plus the context around it: whether the work was done in one sitting, whether the text was pasted in, and how an AI checker classifies the answer as human, AI generated or mixed. None of that is a verdict. A flag is a reason to ask a follow-up question, never a reason to reject someone automatically, and Testlify's own product says so in the interface: AI scores and insights are for guidance only, and human judgment makes the final call.

What qualifications should a translator have?

Translators typically need at least a bachelor's degree to enter the occupation, according to the same federal occupational data cited above, but the degree is rarely the discriminator. Credentials in this field are uneven by language: widely-taught pairs have certification bodies and recognized university programs, while the long tail has neither. Weighting years of experience heavily is a mistake the research argues against. A 2022 meta-analysis of selection methods places structured interviews among the strongest predictors of job performance and puts years of education and general work experience among the weaker ones.

So treat credentials as a filter for risk, not a proxy for skill. A certification tells you someone met a standard once. A scored work sample tells you what they can do this month. Run both where certification exists, and lean on the work sample where it doesn't.

Score that sample with a rubric you wrote before reading the first submission. A meta-analysis of scoring methods found that mechanically combining assessment information predicts outcomes at least as well as experts recombining the same information by judgment. In practice: agree the weightings for accuracy, register, terminology and formatting up front, mark each submission against them, and resist the urge to re-rank the shortlist on feel afterwards.

Translator or interpreter: which do you need?

Translators work with written material. Interpreters work with spoken and signed communication, in real time. The distinction is in the federal occupational definition, and it matters at hiring time because the skills barely overlap: a superb translator may freeze in a live consecutive setting, and a fluent interpreter may write clumsily.

Test in the mode the job actually uses. If the role is 80 percent written localization with occasional meeting support, weight the written work sample accordingly and keep the live element short.

Interpreter interview questions to ask instead

If the role is spoken, swap the take-home for a live exercise and ask about note-taking systems, how they handle a speaker who won't pause, what they do when they miss a phrase, and where they draw the line on summarizing. A listening comprehension assessment gives you a comparable baseline across candidates before anyone books a live session.

What if nobody on the panel speaks the language?

This is the normal case, not the edge case, and it is the reason so many translator hires come down to a gut call. You cannot judge the output, so judge everything around it: the reasoning, the process, the documentation, and a second opinion you pay for.

When you cannot interview or translate in the target language

Four things still work when the panel is monolingual. Ask the candidate to annotate their own translation, explaining three decisions in a language the panel reads; weak candidates cannot explain choices they didn't make. Run a back-translation by a second freelancer and compare meaning, not wording. Give a deliberately ambiguous source sentence and score the clarifying question rather than the output. And use a standardized language proficiency test for the level check, so at least one number on the scorecard is comparable across candidates.

Testlify's real-time translation of test questions helps here too, though the product is honest about its limits: quality varies by language pair, so use it to widen access to the test, not to mark the translation itself.

How do you hire for a rare or low-resource language?

There is no certification, no test vendor and often no second speaker to review the work. UNESCO's endangered-language data puts 1,749 languages, 76 percent of the endangered ones with population data, at fewer than 10,000 speakers, and counts 18 languages spoken by a single person. At that end of the scale, the hiring process has to change shape entirely.

Rare language translator interview: Dura, Kaixana and Kawishana since 2010

Dura in Nepal, and Kaixana and Kawishana in the Brazilian Amazon, sit at the extreme end of that long tail, where the documentation records a handful of speakers or none at all. The picture keeps moving: UNESCO notes that since 2010 more endangered languages are being discovered, so the reference works that a recruiter would check are never quite current. Nobody runs a proficiency test for these pairs. If an NGO brief, a research project or a government contract needs one, the candidate pool may be a handful of people, some of whom are the last documented speakers. Assess on documented fieldwork, community standing and a paid pilot task, and accept that a second reviewer may not exist.

Cawishana translator interview: one language, two spellings

Spelling is its own trap. Kawishana appears in older records as Cawishana, and a Cawishana interview translator search returns a different, smaller set of names than the same search under the modern spelling. Before concluding that nobody exists, run the search under every documented variant of the language name, then check the academic literature for the researchers who have worked on it. They usually know who the speakers are.

1. How did you learn this language, and where did you last use it professionally?
2. Who else works in this pair, and who would you ask to review your work?
3. What written standard or orthography do you follow, and who set it?
4. How do you handle concepts the target language has no word for yet?
5. What documentation exists for this language, and which parts do you trust?
6. How would you price and scope a pilot before we commit to a full project?
7. What is the risk to this project if you become unavailable?
8. Who in the community should review this translation before it is published?

Hire a translator with evidence, not a gut call

Pick the two competencies that decide the role, write the rubric, then collect the evidence. Testlify's test library covers the standardized parts of that: a CEFR English test for level, a grammar and vocabulary test for target-language precision, and open-ended questions with AI-assisted scoring and human override for the work sample itself. To see how the post-editing task and the integrity signals fit your own pipeline, book a demo.

Key takeaways

• Demand is flat, so quality sorting is the whole job. With 2 percent projected growth and most openings coming from replacement, you are not racing other employers for scarce talent. You're choosing well from a steady pool, which means the cost of a loose process lands on quality, not on speed.
• A plain translation test measures the wrong thing now. Machine output is fluent, so fluency no longer proves comprehension. A post-editing task, where the candidate fixes a machine draft and defends the edits, exposes the judgment you're actually buying.
• Every competency needs a named artifact. If you cannot say what a question produces, drop it. The annotation, the terminology note and the clarifying question are all collectable evidence; "strong attention to detail" is not.
• Write the rubric before you read the submissions. Structured scoring holds up better than a panel re-ranking the shortlist afterwards, and it also gives an unsuccessful candidate a real answer about why.
• A monolingual panel can still run a rigorous process. Annotation, back-translation, an ambiguous source sentence and a standardized level test give you four comparable signals without anyone on the panel reading the target language.
• Rare pairs need a different process, not a lower bar. Where no test or second reviewer exists, a paid pilot and community review carry the weight that certification carries elsewhere, and the spelling of the language name is worth checking before you conclude nobody is available.

FAQs