What the work involves
You receive batches of short Hebrew audio clips and score them against a rubric: does the speaker sound native, where does the prosody break, is the pronunciation consistent with a real regional variety or does it drift into something no Hebrew speaker would produce? Some clips are human recordings; many are model output, and the interesting failures are subtle — a correct phoneme inventory with wrong stress placement, unnatural pausing mid-construct chain, flattened intonation on questions, or Modern Hebrew realizations mixed inconsistently with liturgical or Mizrahi pronunciation.
Every rating needs a written justification in English. This is the part that separates useful annotators from fast ones: coordinators need to read your note and understand exactly which acoustic or phonological feature drove the score, in language specific enough that another reviewer would land in the same place. You will also flag ambiguous guideline cases back to project coordinators rather than guessing, since consistency across the annotator pool matters more than any single verdict.
What the screen looks for
- Verifiable native-level Hebrew — micro1's AI interviewer probes spoken fluency directly and follows up on regional features, register shifts, and how you'd describe them.
- Analytic vocabulary, not just intuition — "it sounds off" is not a rating; backgrounds in phonetics, linguistics, voice work, speech therapy, audio engineering, translation, or language teaching help here.
- English at B2 or above for written feedback and interview conversation.
- Rubric discipline — whether you can apply someone else's criteria, including when you personally disagree with them.
Logistics
Fully remote, contractor engagement, asynchronous within task deadlines. Volume is project-driven and fluctuates; most contributors treat it as part-time alongside other work. You'll need quiet listening conditions and decent headphones — laptop speakers are not adequate for judging intonation and articulation detail. Pay bands of $30–65/hr are as observed on the platform and vary by specialization and project, not guaranteed.