Note: The job is a remote job and is open to candidates in USA. Remoat Teams is hiring speech and audio AI experts to inform the design of high-quality audio datasets. The role involves identifying gaps in current models, advising on training and evaluation data, and providing insight into how AI labs and voice AI companies source and assess audio data.

Responsibilities

• Identify where current speech models break down, such as accents, low-resource languages, code-switching, noisy audio, overlapping speakers, and domain-specific vocabulary
• Advise on what training and evaluation datasets are most valuable and how they should be structured
• Help us understand how AI labs and voice AI companies source, annotate, and quality-check audio data
• Provide input on dataset design, quality standards, and customer needs

Skills

• Hands-on experience in one or more of the following: speech-to-text (ASR), speech-to-speech, text-to-speech, or audio model training and evaluation
• Background as a research scientist, ML engineer, or applied researcher at an AI lab, voice AI startup, or tech company
• A strong understanding of where current audio models perform well and where they don't
• Experience working with or selling into AI labs
• Experience sourcing, evaluating, or purchasing training data
• Multilingual ability or experience with non-English speech data

Company Overview