Lead ML Engineer - Audio Evaluation & Metrics

Cantina; United States; Remote; USD 140,000 - 210,000; Full time

Job summary

Cantina Labs seeks an experienced Machine Learning Engineer to focus on audio model evaluation, specializing in speech generation and recognition. You will design comprehensive evaluation pipelines for development and production and build automated dashboards to report results.

As a founding member of the evaluation team, you will lead efforts, collaborate with model, data, and product teams, and help shape the future growth of Cantina's evaluation capabilities.

Qualifications

• Proven ability to design metrics that capture model performance.
• Experience designing user studies on platforms like Mechanical Turk.
• Hands-on experience training and fine-tuning models for evaluation.
• Strong statistical knowledge for comparing results and supporting decisions.
• Excellent software engineering and programming abilities.
• Experience training ASR and TTS models.
• Experience on ML teams handling large-scale datasets and models.

Responsibilities

• Design evaluation pipelines for development and production environments.
• Create user studies to evaluate subjective model aspects.
• Translate requirements into measurable evaluation metrics.
• Develop automated dashboards to monitor model performance and compare results.
• Train new models to capture evolving evaluation metrics.
• Collaborate with the model team to improve models based on results.
• Work with data teams to determine data needed to improve performance.
• Coordinate with product managers to ensure correct measurement of requirements.
• Help grow and lead the evaluation team as a founding member.

Skills

Metric design

User studies MTurk

Model training

Statistical analysis

Engineering skills

ASR training

Large-scale ML

Job description

Cantina Labs seeks an experienced Machine Learning Engineer to focus on audio model evaluation, specializing in speech generation and recognition. You will design comprehensive evaluation pipelines for development and production and build automated dashboards to report results.

As a founding member of the evaluation team, you will lead efforts, collaborate with model, data, and product teams, and help shape the future growth of Cantina's evaluation capabilities.