Speech & Audio ML Research Scientist

Openkrill; United States; Remote; USD 160,000 - 185,000; Full time

Job summary

Innodata is hiring a Research Scientist to own the science behind speech and audio data for ASR, TTS, and related models. You will partner with customers and frontier labs to define data specifications, transcription conventions, and evaluation criteria, going beyond word error rate to measure semantic accuracy and robustness.

You will design experiments, evaluate data choices, and publish benchmarks and methodologies that advance the field and earn trust with clients.

Qualifications

• 5+ years of hands-on industry experience in speech or audio ML.
• Bachelor's degree in computer science, electrical engineering, or a related technical field is required; MS/PhD preferred.
• Trained and evaluated speech or audio models with strong PyTorch fundamentals.
• Fluency with ESPnet, NeMo, SpeechBrain, Kaldi, HuggingFace, and related metrics (WER/CER).
• Experience with multilingual, accented, dialectal, low-resource, or code-switched speech and synthetic/augmented audio.

Responsibilities

• Define how Innodata designs, structures, and evaluates audio data for speech and audio models; validate experimentally.
• Translate model requirements into concrete data specifications (modalities, transcription schemas, sampling).
• Build evaluation methods beyond WER to assess semantic accuracy, robustness, diarization, naturalness, and latency.
• Determine data structuring and sampling across languages, accents, and conditions for model objectives.
• Turn specifications into operational plans with annotation teams and engineering partners.
• Publish benchmarks and methodologies to advance the field and customer trust.

Skills

Speech ML experience

PyTorch

Multilingual speech

Evaluation metrics

Education

Bachelor's in CS/EE

MS/PhD preferred

Tools

ESPnet

NeMo

SpeechBrain

Kaldi

HuggingFace

Forced Alignment

Job description

Innodata is hiring a Research Scientist to own the science behind speech and audio data for ASR, TTS, and related models. You will partner with customers and frontier labs to define data specifications, transcription conventions, and evaluation criteria, going beyond word error rate to measure semantic accuracy and robustness.

You will design experiments, evaluate data choices, and publish benchmarks and methodologies that advance the field and earn trust with clients.