What the work actually is

Audio expert briefs on micro1 vary by client, but they cluster into a few recognisable shapes. You may be listening to text-to-speech output and scoring it on naturalness, prosody, pronunciation of proper nouns, and artifacting — then writing a short justification that a model trainer can act on. You may be comparing two speech-recognition transcripts against reference audio and marking substitution, deletion, and insertion errors, including the ones a WER script scores as correct but a human hears as wrong. You may be labelling acoustic events, tagging speaker turns and overlap in messy multi-speaker recordings, or judging music and sound-design generations for timbre, mix balance, and whether the output matches the prompt. Some tasks are red-teaming: probing an audio model until it mispronounces, hallucinates speech that is not present, or degrades on accented input.

The common thread is that you hear things a metric does not, and you can say why in writing. A rating with no rationale is close to worthless to the client; a rating with a precise, reproducible rationale is the whole product.

What the screen looks for

micro1 runs an AI-led interview before any project match. Expect it to push on where your ears were trained — studio work, broadcast, post-production, speech pathology, linguistics/phonetics, acoustics engineering, localisation QC, or hearing science all count, but the screen wants the specific setting and the specific gear. It will probe calibration: whether you can apply a rubric consistently rather than defaulting to personal taste, what you do when two clips are genuinely tied, and how you handle a guideline you think is wrong. It also checks listening environment and equipment, since ratings made on laptop speakers in a noisy room are not usable data.

Logistics

  • Remote, contract, invoiced hourly or per task; no employment relationship implied.
  • Largely asynchronous, with occasional live calibration sessions on client time zones.
  • Volume is project-driven and lumpy — a heavy month can be followed by nothing.
  • You are expected to supply your own closed-back monitoring headphones or monitors and a quiet space; some briefs also ask for a clean recording setup.
  • Pay band as observed on the platform: $50–100/hr, set per project and by seniority, not guaranteed.