Innodata
Principal Speech Data Linguist
$160K–$185K
Remote - United States +1 · Remote · Principal · Full-time · Posted Sep 24, 2026
- Location
- Remote - United States
North America - APR - Experience
- Typically 8+ years in transcription, segmentation, and speech-data quality.
- Education
- Bachelor's degree in linguistics, phonetics, computational linguistics, or a closely related field; an advanced degree is preferred.
About the job
Own linguistic standards, quality frameworks, and human-in-the-loop workflows for large-scale speech transcription and segmentation, partnering with researchers and customers on model training and evaluation needs.
What you'll do
- Define transcription and segmentation standards; establish quality rubrics, error taxonomies, and adjudication processes.
- Manage deliverable quality end to end, address difficult speech cases, and train and mentor transcribers and reviewers.
- Design human-in-the-loop workflows and collaborate with researchers and customers on how annotation choices affect speech models.
What they ask for
- Bachelor's degree in linguistics, phonetics, computational linguistics, or a closely related field.
- Fluency in phonetic transcription and IPA, with hands-on acoustic and phonetic analysis experience.
- Deep experience with audio segmentation, timestamping, speaker labeling, and diarization.
- Hands-on fluency with speech tools and scripting, including ASR engines, forced alignment, annotation tools, and Python.
Nice to have
- Advanced degree.
- Low-resource language experience; responsible-AI knowledge related to speech is a bonus.
Pay
$160,000 - $185,000 p/year (USD), base salary
Tools
IPA, Praat, Whisper, AssemblyAI, Deepgram, Rev, Speechmatics, Montreal Forced Aligner
AI focus
Sets standards and human review workflows for speech data used to train and evaluate ASR, TTS, and speech and content-understanding models; considers how annotation choices affect model outcomes.
Model-written summary · the original posting is authoritative