Moonlight MCQA - Chinese (Traditional) Generalist

$40–49/hrrate as listed

Write original general-knowledge multiple-choice questions in Traditional Chinese, each with ten answer options and a worked solution, for use in an AI evaluation dataset. The work is remote and self-paced, with pay observed in the $39.69–48.51/hr range.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You draft exam-style multiple-choice questions in Traditional Chinese, pulled from your own academic and professional knowledge rather than from published question banks. Each item needs a clear stem, ten answer options — one defensibly correct, nine plausible distractors — and a written solution that explains why the key is right and why the near-misses fail. Items are expected to be original: copying or lightly paraphrasing textbook, exam, or online questions is the fastest way to fail review.

Day to day this looks less like teaching and more like item writing under constraints. You will pick a topic you can defend, calibrate difficulty so a strong model does not solve it trivially, avoid ambiguity that lets two options both be argued, and keep the language natural Traditional Chinese as used in Taiwan or Hong Kong — not translated-feeling text. Reviewers typically send items back for unclear stems, distractors that are obviously wrong, or solutions that assert rather than reason.

What the platform screens for

Credentials: master's or PhD, plus 2–6 years of professional experience.
Regional and language fit: educated in or based in Taiwan or Hong Kong, with native-level Traditional Chinese.
Domain depth under follow-up: the screen probes whether you can go two or three questions deep on a subject you name, not just describe it.
Evaluation judgment: whether you understand what makes a question hard for a model versus merely obscure, and whether you can spot an ambiguous item.

Logistics

Fully remote and asynchronous. Hours are flexible, with no fixed shifts; most contributors set their own weekly volume, and throughput matters more than time-of-day availability. Onboarding usually includes a paid or unpaid calibration batch where your first items are reviewed closely before you scale up. Pay in the $39.69–48.51/hr band reflects observed rates on the platform and is not guaranteed for any individual contributor.

What they look for

Master's or PhD in any field, with 2–6 years of professional experience
Native-level Traditional Chinese (Taiwan or Hong Kong usage)
Educated in or currently based in Taiwan or Hong Kong
Exam or assessment item writing, including plausible distractor design
Ability to write clear step-by-step worked solutions
Self-directed remote work with consistent weekly output

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Do you hold a master's degree or PhD from an accredited institution?
Master's degree or PhD · required
Do you have between 2 and 6 years of professional experience after completing your undergraduate degree?
Professional experience: 2 to 6 years · required
Is Chinese (Traditional) your native language, or have you lived and worked professionally in Taiwan or Hong Kong for at least 2 cumulative years?
Native or near-native Chinese (Traditional) · required
Were you educated in Taiwan or Hong Kong, or do you currently reside in one of those regions?
Education or residency in Taiwan or Hong Kong · required

0 of 4 answered

Included with this listing

Interview prep — 12 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether the candidate can identify knowledge depth suitable for postgraduate assessment and construct psychometrically sound distractors—not just correct answers.
A strong answer: The candidate names a specific domain (history, economics, science), sketches the core concept they'd test, and describes the reasoning behind 2–3 distractors: common misconceptions, near-correct interpretations, or conceptually related but wrong answers.
Where people lose it: Naming a topic but offering no sense of how distractors are built. Saying 'I'd make ten options' without showing the work of what makes each one defensible to a test-taker.