Description
Deepgram is seeking a Latent Space Models Researcher to pioneer new approaches for scalable, contextualized voice AI. The role focuses on neural audio codecs, steerable generative speech models, interpretable latent embeddings, large-scale synthetic audio data, multimodal speech-to-speech systems, and efficient model architectures and inference. Candidates should have strong foundations in statistical learning, foundation model architectures, data pipelines, controlled experimentation, hardware-aware optimization, and a record of open-source contributions or research publications in speech or language AI.
