Research Staff, Voice AI Foundations

Deepgram

. Pioneer Latent Space Models to address fundamental data, scale, and cost challenges in robust, contextualized voice AI .

Posted 10/11/2026full-timeRemote • NetherlandsLeadWebsite

Core Competencies

Role fit

Core Competencies

Use this summary to align your resume positioning with the role.

Demonstrates expertise in building and optimizing neural audio codecs and generative models for voice AI, with a strong foundation in statistical learning theory and experience in managing large-scale data pipelines. Proven ability to conduct rigorous experiments and contribute to open-source projects in speech processing.

Highest-signal resume keywords

Neural Audio Codec DevelopmentStatistical Learning TheoryMultimodal LearningData Pipeline ManagementOpen-Source Contributions

ATS Keywords

Tailor your resume

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills

Latent Space ModelsGenerative ModelsSpeech ProcessingModel Architecture DesignTraining Scheme DevelopmentInference Algorithm OptimizationReal-Time InferenceControlled Experiment DesignMathematical FormulationAI Automation

Soft Skills

CollaborationCritical ThinkingProblem SolvingVisionary Thinking

Tools & Technologies

Bare-Metal HardwareHigh-Fidelity ReconstructionLow-Bit-Rate CompressionSpeech-To-Speech Systems

Industry Keywords

Speech AISelf-Supervised LearningMultimodal SystemsSTT/TTSEmpathic Responses

About the role

Key responsibilities & impact

Pioneer Latent Space Models to address fundamental data, scale, and cost challenges in robust, contextualized voice AI
Build next-generation neural audio codecs with extreme low-bit-rate compression and high-fidelity reconstruction
Develop steerable generative models synthesizing diverse human speech, including emotional, multi-speaker, noisy, and overlapping-speech scenarios
Create embedding systems that factorize latent spaces into speaker, content, style, environment, and channel dimensions
Use latent recombination to generate synthetic audio data at previously impossible scales
Train multimodal speech-to-speech systems for universal understanding and empathic, human-like responses
Design model architectures, training schemes, and inference algorithms optimized for bare-metal hardware
Enable cost-efficient training on billion-hour datasets and real-time inference for hundreds of millions of concurrent conversations
Conduct rigorous controlled experiments, ablations, evaluations, and stress tests
Collaborate through research publications and open-source contributions

Requirements

What you’ll need

Strong mathematical foundation in statistical learning theory, particularly for self-supervised and multimodal learning
Deep expertise in foundation model architectures and scaling training across multiple modalities
Ability to derive novel mathematical formulations and implement them efficiently
Demonstrated ability to build and maintain massive, high-quality, diverse data pipelines
Experience designing controlled experiments to isolate architectural impacts and validate theoretical insights
Experience optimizing models for real-world deployment, including hardware constraints and efficiency techniques
Track record of open-source contributions or research publications advancing speech/language AI
Ability to identify critical experiments that validate or disprove ideas quickly
Vision to scale successful proofs-of-concept 100x
Strong AI automation and augmentation mindset
Scholarly literature or academic publications related to speech processing, STT/TTS, or similar requested in the application
Legal authorization to work in the country where the role is located
Ability to address visa sponsorship requirements for the country where the role is located

Benefits

Comp & perks

Remote work arrangement
AI-first work environment with opportunities to use and experiment with advanced AI tools
Opportunity to work on transformative voice AI research
Opportunity to contribute to open-source projects and research publications
AI Notetaker interview recording is optional; opting out does not impact candidacy