Sonic Audit Specialist - Japanese

Overview

Mercor is hiring native and near-native Japanese speakers to audit AI training data in Japanese. This is a fully remote independent contractor role with work completed on your own schedule. The work centers on careful listening and written judgment rather than high-volume production, and includes a short calibration phase before production work begins. Payment is weekly on Stripe or Wise based on services rendered.

You will work across two Amazon Sonic audit collections: Multilingual Transcription Audit and Forced Alignment Audit. In both collections, you review audio, compare it against transcriptions or segmented word-level timestamps, apply the relevant audit standards, and provide a short written rationale for each judgment.

What You’ll Do

• Listen to audio recordings of human or agent speech and evaluate whether a transcription accurately reflects what was said.
• Apply a fixed set of error codes, assign pass/fail outcomes, and write concise reasons for each decision.
• Review word-level audio segments and verify that timestamps match the actual start and end of each spoken word.
• Correct boundary issues by splitting, merging, adding, or removing segments when the automated system is inaccurate.
• Tag every applicable error code when a problem is present, and explain the judgment clearly in plain English.
• Complete a short calibration phase before moving into production work.

Requirements

• Native or near-native command of Japanese as spoken in Japan.
• Either being born and raised in a region where this variety is dominant, or having five or more years of residence there with full fluency.
• Strong English reading and writing skills, since conventions, error codes, and rationales are written in English.
• Ability to apply a written rubric consistently across many judgments.
• Careful listening skills, including the ability to distinguish similar sounds and identify word boundaries in continuous speech.
• Ability to tell genuine acoustic ambiguity apart from a clear annotation error.

Preferred Qualifications

• Prior transcription, subtitling, localization, linguistic annotation, or QA experience.
• Prior AI data annotation or model evaluation experience.
• Background in phonetics or linguistics.
• Experience using audio editing or waveform tools.

Additional Details

• This is an independent contractor engagement.
• The role is fully remote.
• Projects may be extended, shortened, or concluded early depending on needs and performance.
• Mercor states that the work will not involve access to confidential or proprietary information from any employer, client, or institution.
• Mercor notes that H1-B and STEM OPT candidates cannot be supported at this time.