Japanese Audio Annotation QA Specialist

About OpenTrain

OpenTrain is the leading platform for people building careers in AI training and data labeling. We help freelancers find and grow steady, skill-matched opportunities, build a unified proof-of-work profile, and turn short-term tasks into durable experience.

OpenTrain AI is the hiring and contracting organization for this role and supports clear expectations, flexible freelance collaboration, and projects that directly improve real AI systems.

About AI training work

AI training (also called data labeling or human feedback work) is the human side of building artificial intelligence. Contributors prepare and evaluate examples — transcriptions, annotations, and quality checks — that models learn from and are evaluated against.

This role sits at the evaluation-grade end of the field: your careful QA helps determine whether multilingual, multi-speaker models meet strict fidelity and annotation standards.

The role

You will review recorded multi-channel group conversations and validate transcription and diarization quality for evaluation-grade conversational datasets. Work focuses on audio fidelity, speaker turns, word-level timestamps, metadata and safety/privacy checks.

This is a fully remote, part-time contractor assignment with a typical expectation of 20+ hours per week.

What you'll do

• Listen to multi-channel audio to detect bleed, clipping, low gain, noise, quiet channels, and other technical fidelity issues.
• Validate verbatim transcription accuracy against strict low-error standards, ensuring spontaneous speech features (overlaps, interruptions, false starts) are preserved.
• Check diarization and turn-level timing, confirm word-level timestamps, and verify correct speaker identification across channels.
• Audit JSON-formatted annotation files and metadata for correctness and consistency.
• Identify and flag safety or privacy issues such as personally identifiable information (PII) or toxic content.
• Provide clear pass/fail judgments with concise, actionable review comments for annotation corrections.

Requirements

• Native proficiency in Japanese required for accurate verbatim transcription and QA.
• Exceptional ear for audio fidelity and attention to detail (bleed, clipping, background noise, gain issues).
• Experience with verbatim transcription and complex multi-speaker conversational dynamics.
• Comfort reading, interpreting, and validating JSON-formatted annotation data.
• Consistent, precise judgment in assessing diarization, timestamps, and speaker labeling.
• Availability for ~20+ hours per week as a remote contractor.

Helpful background (preferred but not required)

• Training or coursework in linguistics or phonetics.
• Professional transcription experience, especially with multi-speaker recordings.
• Experience teaching Japanese or working with spoken-language assessment.