Research Staff, Voice AI Foundations [Remote]
Deepgram · Remote · Australia- Melbourne ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Fully remote role
- Voice AI research
- Latent Space Models
- AI-first culture
- Senior research level
Join Deepgram's research team to develop new paradigms for audio AI, specifically focusing on Latent Space Models. You will work on solving fundamental challenges in data scarcity, computational cost, and model scaling for voice interaction. The team builds production-grade voice agents and APIs used by major organizations worldwide.
What you'll do
- Build next-generation neural audio codecs for low bit-rate compression and high fidelity
- Pioneer steerable generative models to synthesize diverse human speech from latent representations
- Develop embedding systems that factorize codec latent space into interpretable dimensions
- Leverage latent recombination to generate synthetic audio data at massive scales
- Train multimodal speech-to-speech systems for empathic and task-oriented responses
- Design model architectures and inference algorithms optimized for hardware efficiency
What you'll need
- Strong mathematical foundation in statistical learning theory and self-supervised learning
- Deep expertise in foundation model architectures and multi-modal scaling
- Proven ability to bridge theory and practice with efficient implementations
- Experience building data pipelines for massive dataset processing and curation
- Track record of designing controlled experiments to validate architectural innovations
- Experience optimizing models for real-world deployment and hardware constraints
- History of open-source contributions or research publications in speech or language AI
Stack
- Python
- PyTorch
- Triton
- TPU
- Neural Audio Codecs
- Transformers
- SSMs
How you'll work
- Fully remote position
- AI-first mindset required for daily operations
- Fast-paced environment with rapid changes
- Expectation to actively use and build AI tools