Research Staff, Voice AI Foundations - Remote
Deepgram · Remote · The Netherlands ·
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 pioneer new approaches in audio AI, specifically Latent Space Models. You will work on next-generation neural audio codecs, steerable generative models, and embedding systems to enable robust, contextualized voice interactions at scale.
What you'll do
- Build neural audio codecs for low bit-rate compression and high fidelity
- Develop steerable generative models for diverse human speech synthesis
- Create embedding systems to factorize codec latent space dimensions
- Generate synthetic audio data via latent recombination at scale
- Train multimodal speech-to-speech systems for empathic responses
- Design model architectures optimized for hardware efficiency and real-time inference
What you'll need
- Strong foundation in statistical learning theory and self-supervised learning
- Deep expertise in foundation model architectures and multi-modal scaling
- Ability to derive mathematical formulations and implement them efficiently
- Experience building pipelines for massive dataset processing and curation
- Track record of designing controlled experiments to validate innovations
- Knowledge of hardware constraints and model optimization for deployment
- History of open-source contributions or publications in speech/language AI
Stack
- Python
- PyTorch
- Triton
- TPU
- Neural Audio Codecs
- VQ-VAE
- Transformers
- SSMs
How you'll work
- Fully remote position
- AI-first operating rhythm
- Rapid pace of change and innovation
- Expectation to use advanced AI tools daily