Research Scientist: Model Evaluation & Quality Metrics

Sanas; Palo Alto (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Sanas in Palo Alto is seeking a Research Scientist focused on rigorous evaluation of speech AI models. You will define meaningful metrics, build scalable evaluation pipelines, and align research progress with product impact across Accent Translation, Noise Cancellation, and Speech Enhancement.

The role requires deep expertise in evaluation methodologies, strong programming in Python and PyTorch, and the ability to run human studies and ablations at scale, partnering with ML research, product,

Qualifications

• 4+ years of research or applied research experience in speech, audio, or NLP.
• Experience designing and running human evaluation studies with statistical rigor.
• Strong engineering skills to build production-quality evaluation pipelines in Python and PyTorch.

Responsibilities

• Design and own evaluation frameworks across Sanas's model portfolio, ensuring metrics capture meaningful progress.
• Develop quantitative metrics for subjective and perceptual qualities and align with product goals.
• Build evaluation systems bridging automated metrics and human judgment; run robust listening tests.
• Define evaluation splits and benchmark suites reflecting real production conditions.

Skills

Research experience

Python

PyTorch

Speech/Audio/NLP

Evaluation methodology

Job description

Sanas in Palo Alto is seeking a Research Scientist focused on rigorous evaluation of speech AI models. You will define meaningful metrics, build scalable evaluation pipelines, and align research progress with product impact across Accent Translation, Noise Cancellation, and Speech Enhancement.

The role requires deep expertise in evaluation methodologies, strong programming in Python and PyTorch, and the ability to run human studies and ablations at scale, partnering with ML research, product,