Member of the Research Staff (AI)

Location: Remote from Melbourne, Australia

Company

Deepgram provides real-time voice AI APIs and self-hosted software for speech-to-text, text-to-speech, and production-grade voice agents.

What you will do

• Develop latent space models that address data, scale, and cost challenges in voice AI.
• Build neural audio codecs with extremely low bit-rate compression and high-fidelity reconstruction across diverse audio.
• Develop steerable generative models for expressive, multi-speaker, and noisy speech synthesis.
• Create embedding systems that factorize speaker, content, style, environment, and channel characteristics.
• Use latent recombination to generate synthetic audio datasets and train multimodal speech-to-speech systems.
• Design hardware-efficient architectures, training methods, and inference algorithms for large-scale and real-time deployment.

Requirements

• Strong mathematical foundation in statistical learning theory, self-supervised learning, and multimodal learning.
• Deep expertise in foundation model architectures and scaling training across multiple modalities.
• Ability to derive novel mathematical formulations and implement them efficiently.
• Experience building and curating large-scale data pipelines while maintaining quality and diversity.
• Experience designing controlled experiments and optimizing models for real-world hardware constraints.
• Open-source contributions or research publications advancing speech or language AI.

Nice to have

• Experience with neural audio codecs, discrete representation learning, generative speech models, or large-scale speech datasets.
• Experience with efficient architectures, custom kernels, quantization, or real-time and on-device inference.

Culture & Benefits

• AI-first working style with active use and experimentation with advanced AI tools.
• Fast-moving research environment requiring adaptability, creativity, and continuous learning.
• Focus on validating ideas through critical experiments and scaling successful proofs of concept.
• Commitment to impactful, scalable, and reproducible research in voice AI.