What the work involves
You draft exam-style multiple-choice questions in Modern Standard Arabic, pulled from areas you can actually defend: your degree subject, your professional field, and general knowledge grounded in Arabic-language sources. Each item needs ten answer options — one defensibly correct, nine distractors that are plausible enough to be tempting but wrong for a reason you can articulate — plus a written solution that walks through why the key holds and why the near-misses fail.
The difficulty target is a model that already answers easy questions correctly. That means most of your effort goes into distractor craft and into avoiding the failure modes reviewers reject: ambiguous keys, two arguably-correct options, questions answerable by elimination or surface pattern-matching, and items that are really translated English trivia rather than content a well-educated Arabic speaker would recognize as native. Expect revision requests on early batches while you calibrate.
What the platform screens for
- Verifiable credentials: master's or PhD, and 2–6 years of professional experience the screen can probe in specifics.
- Native-level Arabic plus education or residence in an Arabic-speaking country — you'll be asked about register, dialect handling, and terminology choices.
- Domain depth under follow-up: the AI interviewer pushes past your first answer to see whether the expertise is real.
- Evaluation judgment: whether you can explain what makes a distractor good, and when a question is unsalvageable.
Logistics
Remote, asynchronous, no fixed shifts. Work is assigned in batches with deadlines rather than scheduled hours; contributors commonly commit 10–20 hours per week, though volume varies with project demand. Pay is hourly within the observed band above — rates depend on assessed expertise and are not guaranteed. Some projects pay per accepted item instead, so read your specific assignment terms.