Latest AI JobsBrowse Jobs
Deepgram logo

Research Staff, Voice AI Foundations - Remote

DeepgramThe NetherlandsResearch
Apply on company site
Salary
Not listed
Job type
Full-time
Work mode
Remote
Posted
10 Oct 2026
Status
Open · checked 7d ago

About this role

Research Staff, Voice AI Foundations - Remote

A Research Staff position at Deepgram, working remotely from the Netherlands, is suited for experts in audio AI who want to develop new latent space models. The role focuses on creating advanced neural audio codecs, generative models, and embedding systems for voice applications.

Role Overview

Deepgram is building foundational technology for the Voice AI economy, offering real-time speech-to-text and text-to-speech APIs. The Research Staff member will lead investigations into Latent Space Models (LSMs) to address data scarcity, high dimensionality, and cost challenges in large-scale audio AI.

Key Responsibilities

Design and implement next-generation neural audio codecs that compress audio to very low bit-rates while preserving high fidelity.
Create steerable generative models capable of synthesizing diverse speech styles, emotions, multi-speaker dialogue, and noisy environments.
Develop embedding systems that separate codec latent space into interpretable dimensions such as speaker identity, content, style, and environment.
Use latent recombination techniques to produce synthetic audio at massive scale, enabling new joint model-data scaling approaches.
Explore multimodal speech-to-speech systems that can understand any human speaker regardless of acoustic conditions.

Requirements

Demonstrated AI-first mindset; regular use and experimentation with advanced AI tools is expected.
Ability to quickly adopt new models and integrate AI into daily workflows.
Comfort with rapid change and willingness to experiment, adapt, and learn continuously.
Strong research background in audio processing, deep learning, or related fields.

Nice to Have

(No additional preferences listed in the posting.)

Skills and Tools

Deep learning for audio (e.g., neural codecs, generative models)
Knowledge of latent space representations and factorization techniques
Experience with large-scale audio datasets and synthetic data generation

About the Company

Deepgram provides cloud APIs and on-premises software for speech-to-text, text-to-speech, and production-grade voice agents. Over 200,000 developers and more than 1,300 organizations, including Twilio and Cloudflare, use its platform. The company has processed more than 50,000 years of audio and transcribed over 1 trillion words. Backed by a Series C round, Deepgram emphasizes an AI-first operating rhythm where every team member actively applies advanced AI tools.

How to Apply

Candidates should apply through Deepgram’s official careers page.

Skills and tech stack

neural audio codecslatent space modelsgenerative speech modelsembedding systemsLLMsTransformersDeep LearningDiffusion

Questions candidates ask

Is this position fully remote?

Yes, the role is listed as Remote and the location is the Netherlands.

What type of research will I be doing?

The research focuses on Latent Space Models, neural audio codecs, steerable generative speech models, and embedding systems for voice AI.

What mindset does Deepgram expect from its employees?

Deepgram expects an AI-first mindset where employees regularly use and experiment with advanced AI tools and integrate AI into their daily work.

Do I need prior experience with large-scale audio data?

Experience with large-scale audio processing, deep learning for audio, or related research is required.

Are there any specific educational requirements?

The posting does not specify any education requirements.

Ready to apply?

The application is on the official Deepgram careers page.

Apply on company site