We are looking for experienced AI data annotators to label audio recordings from 11 sourced voices in the following languages: Spanish, Italian, Swedish, Basque, Galician, Portuguese, Catalan, Arabic, and Japanese. The recordings will be used to train AI TTS and LLM-related systems, so annotation quality and consistency are extremely important. Annotators will follow detailed written instructions and label audio files according to our internal guidelines. The work will include audio quality review, text/audio validation, issue removal, change of turn labeling, EOS validation, and consistency checks. You should have experience with AI data annotation, audio labeling, and speech-data workflows. Experience with transcription QA, TTS dataset annotation, voice data labeling, and/or LLM evaluation or assessment is preferred. You must demonstrate strong attention to detail, the ability to follow precise written instructions, and comfort using an annotation platform. Reliable communication, consistent turnaround, and native and/or fluent understanding of the target language are required. You should be able to incorporate QA feedback and complete rework when requested. Annotators will add notes where needed, flag problematic recordings directly to employer such as unclear/noisy/incomplete audio, and pass a short onboarding annotation call before being assigned production tasks.
$270
$270
Flexible
< 1 month
13
Audio recording Annotations for TTS/LLM training.
Software
Hiring Type
Required Location
Workload / Schedule
Weekly commitment can be adjusted based on throughput targets. Project duration for each audio recording is expected to take up to 10-25 hours. Labelers should follow milestone deadlines and quality checkpoints. Depending on quantity of talent language proficiency, annotators may be assigned to 1 or more languages. Each language may have up to two separate voice's in distinct genders, with various smaller recording's to annotate (up to 5).
Software
Data Type
Task Types
Subject Matter / Industry
Language
Proposals: 16
Invites sent: 0
Unanswered invites: 0
Share link