- Produce verbatim transcripts from audio recordings with 98% accuracy.
- Timestamp segments precisely and label speakers in multi-speaker audio.
- Annotate non-speech events and follow quality standards.
Enterprise Client
4 open remote positions
They are a leading provider of high-quality, diverse datasets for AI model development. The company's size and culture are not specified in the posting.
Benefits Overview 1 of 4 jobs
Flexible hours
(1)
Remote work
(1)
Open Positions
- Produce verbatim transcripts from single-speaker and multi-speaker audio recordings with a target accuracy rate of 98% or higher.
- Timestamp transcript segments with precise alignment to audio, maintaining offsets of less than 500 milliseconds.
- Identify and label speakers in conversations involving two to three participants.
- Design and develop coding benchmarks used to evaluate frontier AI models.
- Analyze AI-generated code for correctness, reliability, efficiency, and edge cases.
- Build and maintain scalable data pipelines that support AI evaluation workflows.
- Record short video and/or audio clips with emotional expression.
- Listen and transcribe various video and/or audio clips.
- Act short scripts in different tones or emotional styles.