Sonic Audit Specialist – Mandarin at Mercor

Sonic Audit Specialist (Mandarin Chinese) | $21.50/hr | Worldwide Remote

Mercor is seeking native and near-native Mandarin Chinese speakers to audit AI training data for Amazon's Sonic collections. This is careful, detail-oriented listening and evaluation work — not content production. If you have a strong ear for spoken Mandarin and a disciplined approach to applying written standards, this role is for you.

About the Work

You will work across two Amazon Sonic audit collections, each running at 5.00 audit hours per task:

• Multilingual Transcription Audit: Listen to recordings of human or agent speech, review the transcription produced by an annotator, and judge whether it accurately reflects what was spoken. Apply a fixed set of error codes, assign a pass/fail verdict, and write a concise rationale for your decision.
• Forced Alignment Audit: Review audio segmented at the word level by an automated system. Verify that start and end timestamps correctly match each spoken word, and that the reference transcription matches the spoken form. Correct boundary errors by splitting, merging, adding, or removing segments as needed.

Every judgment requires a written explanation, making this a quality-focused role rather than a high-volume one.

Who We're Looking For

• Native or near-native Mandarin Chinese (mainland China variety) — born and raised in a dominant-speaking region, or fully fluent with 5+ years of residence there. This is a hard requirement.
• Strong English reading and writing skills — all rulebooks, error codes, and audit rationales are in English.
• Disciplined rubric application — ability to apply a written standard consistently across hundreds of judgments. Experience in transcription, subtitling, localization, linguistic annotation, or QA is a strong plus.
• Careful listening ability — skilled at distinguishing similar sounds, identifying word boundaries in continuous speech, and differentiating acoustic ambiguity from clear annotator error.

Bonus Qualifications

• Prior AI data annotation or model evaluation experience
• Background in phonetics or linguistics
• Familiarity with audio editing or waveform tools

What to Expect

You will complete a short calibration phase before production work begins. All submissions require tagging every applicable error code (not just the first) and a plain-language explanation of your verdict, with critical errors clearly identified.