Research Scientist, Speech & Audio (AI)

Location: Remote - United States

Salary: $160,000–$185,000 per year, based on experience, skills, and qualifications.

Company

hirify.global is a global data engineering company that provides data, evaluation frameworks, and human expertise for generative AI systems.

What you will do

• Define data structures, annotation schemas, sampling strategies, and evaluation criteria for ASR, text-to-speech, speech-to-speech, conversational voice, diarization, verification, audio-language, and streaming models.
• Build evaluation methodologies covering semantic accuracy, accent and noise robustness, code-switching, diarization error rate, speech naturalness, intelligibility, and streaming latency.
• Design audio coverage across languages, accents, dialects, acoustic conditions, speaker demographics, emotional and paralinguistic characteristics, and single- or multi-speaker settings.
• Partner with transcription, linguistics, audio solutions, engineering, annotation, and subject-matter teams to turn model objectives into operational data and evaluation programs.
• Fine-tune and evaluate models on hirify.global data, using ablations to connect data decisions with measurable improvements.
• Design adversarial evaluations, document findings, and publish benchmarks, methodologies, and research papers.

Requirements

• Approximately 5+ years of hands-on industry experience in speech or audio machine learning; a current PhD research agenda may offset experience at the lower end.
• Bachelor’s degree in computer science, electrical engineering, or a related technical or quantitative field.
• Hands-on experience training and evaluating ASR, TTS, speech-to-speech, speaker, or audio-language models, with strong PyTorch fundamentals.
• Experience with ESPnet, NeMo, SpeechBrain, or Kaldi, Hugging Face, forced alignment, WER/CER, and advanced speech evaluation metrics.
• Experience with multilingual, accented, dialectal, low-resource, or code-switched speech and synthetic or augmented audio.
• First-author publications or strong open-source contributions at venues such as Interspeech, ICASSP, ASRU, SLT, or NeurIPS.

Nice to have

• Advanced degree in a relevant field.
• Experience with responsible-AI evaluation and red-teaming, including spoofing, voice-cloning robustness, or bias across accents and languages.

Culture & Benefits

• Direct collaboration with customers and frontier AI laboratories developing speech and audio models.
• Work spans research, data design, evaluation methodology, annotation, and engineering implementation.
• Practical experience is valued alongside formal credentials.
• Research findings can be developed into benchmarks, methodologies, open-source contributions, and papers.