Research Engineer - Audio & Speech Models (AI)

Location: San Francisco, United States; on-site

Company

Develops open-source AI models for audio and speech applications.

What you will do

• Build the next generation of open-source audio autoencoders, ASR, TTS, SSL, and speech-to-speech models.
• Run large-scale audio model training and optimize the training stack.
• Collect, process, and evaluate audio datasets.
• Design and improve model architectures and training methodologies through ablations.
• Develop research projects from conception through experimentation and write-up.
• Contribute to large-scale engineering implementation and existing codebases.

Requirements

• Expertise training audio-domain models, including text-to-speech, ASR, speech-to-speech, or speech-emotion-recognition models.
• Experience training audio autoencoders and working with audio signal processing.
• Experience with diffusion models, consistency models, or GANs.
• Experience training on large-scale, multi-node GPU clusters.
• Strong experimental methodology and understanding of highly parallel data processing pipelines.
• Proficiency with PyTorch and Python, plus published machine learning research and a postgraduate scientific degree.

Culture & Benefits

• Research and engineering excellence are valued equally, with a methodical approach to ambitious goals.
• Encouragement to pursue new ideas and move quickly from experimentation to impact.
• Medical, dental, vision, and FSA plans.
• Competitive compensation, 401(k), unlimited PTO, and company holidays.
• Relocation and immigration support may be available on a case-by-case basis.
• In-office snacks and meals in a collaborative, high-energy environment.