The Netherlands
Remote
Senior
Full Time
voice AIresearchmachine learningdeep learningspeech recognitionfoundation modelslatent space modelsneural audio codecsmultimodal learningself-supervised learning
Requirements
- •Strong mathematical foundation in statistical learning theory relevant to self-supervised and multimodal learning
- •Deep expertise in foundation model architectures and scaling training across multiple modalities
- •Proven ability to bridge theory and practice with novel mathematical formulations and efficient implementation
- •Experience building data pipelines for processing and curating massive datasets with quality and diversity
- •Track record of designing controlled experiments to validate architectural innovations and theoretical insights
- •Experience optimizing models for real-world deployment considering hardware constraints and efficiency
- •History of open-source contributions or research publications advancing speech/language AI
What You'll Do
- •Pioneer the development of Latent Space Models (LSMs) for voice AI
- •Build next-generation neural audio codecs for low bit-rate compression and high fidelity reconstruction
- •Develop steerable generative models to synthesize diverse human speech
- •Create embedding systems to factorize codec latent space into interpretable dimensions
- •Generate synthetic audio data at large scale using latent recombination
- •Train multimodal speech-to-speech systems for understanding and producing human-like responses
- •Design model architectures, training schemes, and inference algorithms optimized for hardware and cost efficiency
About Deepgram
Deepgram specializes in providing AI-powered speech-to-text technology that offers audio intelligence, text-to-speech, and voice agent API.
San Francisco, CA
100 - 250
AI & Machine Learning