Machine Learning Researcher Audio

You will research and develop speech-to-text, text-to-speech, large language model, and neural audio codec systems. You will train models at scale, build data and training pipelines, run rigorous experiments, optimize real-time inference, and deploy validated approaches into production.

Responsibilities

• Design and train expressive text-to-speech models
• Develop neural audio codec-based text-to-speech architectures
• Improve prosody, emotional expression, and multi-speaker robustness
• Build and fine-tune robust speech-to-text systems
• Research and implement neural audio codecs
• Curate and process large audio datasets
• Design distributed GPU training pipelines
• Design and evaluate ablation studies
• Optimize production inference for low latency

Requirements

• Experience with self-supervised, multimodal, or generative modeling
• Hands-on experience building or scaling text-to-speech, speech-to-text, or neural audio codec systems
• Familiarity with large-scale speech datasets and audio variability
• Strong understanding of audio quality, prosody, and conversational dynamics
• Experience training and serving large models on modern accelerators
• Knowledge of quantization, kernel optimization, and memory efficiency
• Understanding of real-time telephony or streaming constraints
• Experience designing controlled experiments and ablations

Benefits

• Healthcare
• Dental insurance
• Vision insurance
• Equity
• Tools needed to succeed
• Office in Levi's Plaza with rooftop views