← All roles

Research Staff, Voice AI Foundations [Remote]

Deepgram· Research
Apply on company site
RemoteFull-timePosted 1w ago

About the role

Research Staff role focused on developing Latent Space Models (LSMs) for robust, contextualized voice AI, including neural audio codecs, steerable generative models, latent space factorization, and scalable, hardware-aware architectures for billion-hour datasets and real-time inference.

Responsibilities

  • Pioneer development of Latent Space Models (LSMs) to address data, scale, and cost challenges in voice AI.
  • Build next-generation neural audio codecs with extreme low-bit-rate compression and high-fidelity reconstruction across a world-scale corpus.
  • Pioneer steerable generative models to synthesize diverse human speech from codec latent representations under various conditions.
  • Develop embedding systems to factorize codec latent space into interpretable dimensions (speaker, content, style, environment, channel) and enable latent recombination.
  • Use latent recombination to generate synthetic audio data at scale to enable joint model and data scaling for audio; train multimodal speech-to-speech systems that are robust across demographics and environments and can achieve conversational/task objectives.
  • Design model architectures, training schemes, and inference algorithms optimized for hardware to enable cost-efficient training on billion-hour datasets and real-time inference for hundreds of millions of concurrent conversations.

Qualifications

  • Strong mathematical foundation in statistical learning theory, especially self-supervised and multimodal learning.
  • Deep expertise in foundation model architectures with scalability across modalities.
  • Ability to bridge theory and practice, deriving novel formulations and implementing them efficiently.
  • Experience building data pipelines for massive datasets ensuring quality and diversity.
  • Track record of controlled experiments isolating architectural innovations and validating insights.
  • Experience optimizing models for real-world deployment with hardware-aware efficiency.
  • History of open-source contributions or research publications advancing speech/language AI.

Skills

Required:Statistical learning theoryFoundation model architecturesData pipelines for large-scale datasetsExperimental design

Benefits

  • Not specified in the job description.
Last updated Oct 14, 2026. Interested? Apply on the company site.
Deepgram logo

About Deepgram

Building foundational AI for speech transcription and understanding.

Deepgram is a foundational AI company on a mission to understand human language. We give any developer access to the most advanced speech AI transcription and understanding with just an API call. Our models deliver the fastest, most accurate transcription alongside contextual features like summarization, sentiment analysis, and topic detection. Beyond that, developers can: 🔊 Process live-streaming or pre-recorded audio 🌎 Transcribe in dozens of languages ⚙️ Train custom models for unique use cases 🔑 Access deep NLU with a unified API 💻 Build in any programming language with our SDKs ✅ Deploy on-prem or on DG’s managed cloud 📈 Get scalable GPU infra for training and inference Deepgram is a proud NVIDIA partner and Y Combinator company, and we recently completed a $72M Series B to define the future of AI Speech Understanding, making us the most-funded speech AI company at its stage.

ai-enhanced-learningapiartificial-intelligencebrex-fastest-growing-2025liquid-ai-recruiting-targetspeech-to-texttext-to-speechvoice-ai-stack
Company profileWebsite

More open roles at Deepgram