Software Engineer - Voice Model

You will build speech-data pipelines, train and post-train speech-language models, and evaluate voice-model quality. You will integrate models into real-time applications and support deployment of reliable, low-latency multilingual voice experiences.

Responsibilities

• Design and execute large-scale speech-data curation and processing pipelines.
• Train and post-train speech-language models.
• Build and iterate an evaluation framework for voice-model performance.
• Integrate voice models into applications and real-time environments.
• Define spoken-interaction specifications and support deployment.

Requirements

• Expert Python proficiency for AI/ML systems.
• Experience processing large-scale datasets with Spark and Ray.
• Proficiency pre-training and post-training speech-language models with JAX or PyTorch.
• Experience with supervised fine-tuning and reinforcement learning.
• Ability to run evaluation pipelines with objective metrics, human preference studies, factuality checks, and A/B testing.
• Experience with large-scale distributed training and inference systems on Kubernetes.

Benefits

• Equity
• Medical coverage
• Vision coverage
• Dental coverage
• 401(k) retirement plan
• Short-term disability insurance
• Long-term disability insurance
• Life insurance
• Discounts and perks