Research Staff, Voice AI Foundations [Remote]
Deepgram · London, UK · Remote- Ireland · Switzerland · The Netherlands · Canada · Australia- Melbourne · EMEA - Remote
Applied AIremoteStaff
At a glance
Highlights
- Own Latent Space Models research end-to-end to ship under real-time constraints.
- Build neural audio codecs compressing general audio to ultra-low bitrates.
- Pioneer steerable generative models synthesizing the full range of human speech.
Ideal candidate
A first-principles researcher who thrives on open-ended frontier problems and ships breakthroughs to production. You work AI-natively, reason across domains, and care about making voice the natural interface to technology.
Requirements
- Design controlled experiments and ablations that isolate what works
- Built new foundation-model architectures or scaled existing ones across modalities
- Work AI-native and build tooling, infrastructure or evals to unblock work
- Reason from first principles and explain deep technical work clearly
- Advanced proficiency in Python and PyTorch building models from scratch
- PhD or Master's in ML, CS, math, physics, signal processing or equivalent experience
- remote policy
- Remote
- hiring process
- In interviews, expect to walk through a workflow you built and where you drew the line between AI and your own judgment.
Lab profile
About Deepgram
Real-time voice AI platform with speech models for conversational applications.
What they build
- Voice Agent Platform
- Speech models
- Real-time voice infrastructure
Known for
- Real-time speech processing
- Production-scale conversational voice AI
- Voice foundations and LLM research
Key facts
- 5,000+ organizations in production
Posted · Last seen