Phonetic Linguist - Singing Voice Data Annotation for AI

We are seeking a highly qualified Phonetic Linguist to join our team working on cutting-edge AI voice synthesis technology. This role involves precise annotation of professional singing recordings to create training data for advanced voice synthesis models.

The position requires segmenting singing audio into individual phonemes with millisecond-level accuracy using spectrogram analysis. You will work with recordings across multiple languages, identifying and labeling vocal techniques including breathing patterns, glottal stops, vocal fry, and silence markers.

This is a remote contract position offering the opportunity to work with proprietary annotation technology while contributing to breakthrough developments in AI-powered vocal synthesis. The role involves collaboration with an international team and handling confidential voice synthesis research.

Key responsibilities include

Segmenting singing recordings into 5-30 second phrases with precise timing

Correcting automated phoneme predictions using visual and auditory analysis

Labeling special vocal elements (breathing, silences, glottal stops)

Handling cross-linguistic pronunciation variations in musical contexts

Maintaining consistent quality standards across large datasets

Working with proprietary web-based annotation platforms

Technical requirements

Reliable high-speed internet connection

Professional audio equipment for precise listening

Quiet workspace suitable for detailed audio analysis

Availability for 20-30 hours per week

Comfortable with NDA and confidentiality requirements

Requirements

Essential qualifications

Master's or PhD in Linguistics, Phonetics, or related field

Expert knowledge of International Phonetic Alphabet (IPA)

Proven experience with acoustic analysis software (Praat, ELAN)

Strong background in phonetic transcription and spectrogram interpretation

Experience with time-aligned annotation workflows

Native or near-native English proficiency

Additional language skills (Japanese, Mandarin, Korean, Spanish preferred)

Preferred qualifications

Experience with speech synthesis or voice technology projects

Understanding of vocal techniques and singing styles

Background in computational linguistics or audio signal processing

Previous work with AI training data annotation

Research experience in acoustic phonetics or speech science

Highlights

Work on breakthrough AI voice synthesis technology, annotating singing recordings with millisecond precision for next-generation vocal AI models.

Originally posted on Himalayas