Research Scientist, Speech & Audio

What you'll do

• Define audio data specifications, coverage, transcription schemas, and evaluation criteria for speech and audio models.
• Run model fine-tuning, evaluations, and ablation studies; design adversarial tests and turn findings into improved data and published research.
• Partner with customers, research teams, linguistics and transcription leads, engineers, and annotation teams to deliver data and evaluation programs.

What they ask for

• Bachelor’s degree in computer science, electrical engineering, or a related technical or quantitative field.
• Hands-on experience training and evaluating speech or audio models, with strong PyTorch fundamentals.
• Experience with multilingual, accented, dialectal, low-resource, or code-switched speech, and synthetic or augmented audio.
• Experience building evaluation sets and assessing dataset coverage; clear communication and rigorous, reproducible experimentation and documentation.

Nice to have

• Advanced degree (MS or PhD) in a relevant field.
• First-author publications or strong open-source contributions at venues such as Interspeech, ICASSP, ASRU, SLT, or NeurIPS; responsible-AI evaluation or red-teaming experience is a bonus.

Pay

$160,000 - $185,000 p/year (USD), base salary

Tools

PyTorch, ESPnet, NeMo, SpeechBrain, Kaldi, HuggingFace, forced alignment, WER/CER

AI focus

Research, fine-tune, evaluate, and improve speech and audio AI models through data design, benchmarking, robustness testing, and evaluation methodology.