Latest AI JobsBrowse Jobs
Deepgram logo

Research Staff, Voice AI Foundations [Remote]

DeepgramAustralia- MelbourneResearch
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

Role Overview

This research staff position focuses on Voice AI foundations and is remote, suitable for experienced researchers interested in audio modeling and generative AI.

Key Responsibilities

Design and implement neural audio codecs that provide ultra‑low bitrate compression while preserving high‑fidelity audio.
Create steerable generative models capable of producing diverse speech styles, emotions, and multi‑speaker interactions, even in noisy environments.
Build embedding systems that separate codec latent space into clear dimensions such as speaker identity, content, style, environment, and channel effects.
Use latent recombination techniques to synthesize large volumes of synthetic audio data for scaling model training.
Explore multimodal speech‑to‑speech architectures that can understand and respond to any human voice.
Continuously experiment with emerging AI tools and integrate them into daily research workflows.

Requirements

Strong background in audio signal processing and machine learning research.
Experience developing neural codecs, generative models, or latent space representations for audio.
Ability to factorize high‑dimensional latent spaces into interpretable components.
Proven track record of publishing or delivering research that advances audio AI.
Comfort with rapid iteration, adopting new AI models, and adapting workflows in a fast‑moving environment.

Nice to Have

Experience with large‑scale synthetic data generation for audio.
Familiarity with multimodal speech‑to‑speech system design.

Skills and Tools

Audio modeling and signal processing concepts.
Machine learning frameworks for deep learning (e.g., PyTorch, TensorFlow).
Knowledge of latent space manipulation and generative modeling techniques.

About the Company

Deepgram provides real‑time speech‑to‑text, text‑to‑speech, and voice agent APIs used by over 200,000 developers and 1,300 organizations worldwide. The company processes massive audio volumes, having transcribed more than one trillion words, and operates with an AI‑first culture that expects continuous experimentation and rapid adaptation.

How to Apply

Candidates should apply through Deepgram’s official careers page.

Skills and tech stack

LLMsTransformersDeep LearningDiffusion

Questions candidates ask

Is this role fully remote?

Yes, the position is listed as remote.

What type of research will I be doing?

The work involves developing neural audio codecs, steerable generative models, embedding systems for latent spaces, and large‑scale synthetic audio generation.

Do I need to have experience with specific AI tools?

The posting expects familiarity with advanced AI tools and the ability to adopt new models quickly, but no specific tools are mandated.

Will I be working with Deepgram’s existing voice APIs?

The role focuses on foundational research that underpins Deepgram’s voice AI platform, which includes their speech‑to‑text and text‑to‑speech services.

Ready to apply?

The application is on the official Deepgram careers page.

Apply on company site