Korean Audio Annotation QA Specialist

About OpenTrain

OpenTrain is the #1 platform for people building careers in AI training and data labeling. We help freelancers discover specialized annotation work, consolidate opportunities, and build a single portfolio of AI-training skills and projects they control. Creating an OpenTrain account is free.

About AI training and why it matters

AI training (also called data labeling or annotation) is the human work behind modern AI systems: people prepare, check, and refine examples that models learn from. This role places you at the center of multilingual, multi-speaker audio evaluation—directly influencing how speech and conversation AI behave.

• 100% remote and highly flexible—fit work around your schedule.
• Accessible entry points: many projects require language skill and attention to detail rather than prior AI experience.
• Work on real conversational data and help improve cutting-edge multilingual speech systems.

The role

OpenTrain is hiring a Korean Audio Annotation QA Specialist to review evaluation-grade, multi-channel recordings and the validated transcripts and diarization produced from them. This part-time contractor role requires at least 20 hours per week and is open worldwide.

• Role type: Contractor, Part-time.
• Time commitment: 20+ hours/week.
• Work is fully remote and focuses on high-accuracy QA for audio, transcripts, timestamps, speaker labels, and metadata.

What you'll do

• Listen to multi-channel audio and assess fidelity, quiet recording conditions, channel bleed, clipping, and low gain.
• Validate human-checked transcripts for strict verbatim accuracy and low word error rates.
• Confirm turn-level and word-level timestamps, speaker labels, overlaps, interruptions, and false starts.
• Review diarization and channel isolation to ensure speakers are correctly separated.
• Check metadata, demographic markers, contextual domains, and conversational tags for accuracy.
• Audit sessions for PII, toxic, or sensitive content and provide clear pass/fail or agree/disagree reviews with actionable comments.
• Work with JSON annotation data and provide precise, consistent feedback on annotation quality.

Requirements

• Native proficiency in Korean required for detailed audio and transcript review.
• Strong ear for audio fidelity and subtle recording defects (bleed, clipping, low gain).
• Experience with complex multi-speaker dialogue and verbatim transcription conventions.
• Careful judgment when validating timestamps, diarization, speaker labels, and annotation metadata.
• Comfort reviewing overlapping speech and working with JSON-format annotation files.
• Entry-level applicants are welcome provided they meet the language and audio QA skills above.

Helpful background (not required)

• Training or study in linguistics, phonetics, or speech-language studies.
• Professional transcription or QA experience on multi-speaker audio.
• Previous work reviewing language data with strict accuracy targets.