Director, Text-to-Speech Synthesis Research
Deepgram · Remote · London, UK ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Director level role
- Fully remote
- TTS & Audio AI
- Hands-on leadership
- Production-focused
Lead the end-to-end Text-to-Speech research program at Deepgram, defining technical direction and model architecture. You will balance high-level strategy with deep technical involvement, guiding researchers while personally tackling complex problems in speech generation.
The role requires building a high-performing team and establishing evaluation frameworks that drive measurable production gains. You will partner with engineering and product leaders to ensure research advances translate into deployed, scalable voice AI solutions.
What you'll do
- Own the TTS research roadmap and decide on technical directions for speech quality
- Drive advances in neural audio modeling, prosody, controllability, and multilingual speech
- Design experiments, diagnose model failures, and solve high-leverage technical problems
- Build automated metrics and human perceptual assessments for model evaluation
- Hire and develop researchers and tech lead managers while maintaining high technical standards
- Partner with engineering and product leadership on ship-readiness and external representation
What you'll need
- Deep expertise in modern TTS, speech generation, or audio generative modeling
- Track record of personally training and improving large-scale neural models
- Experience setting research direction under uncertainty and allocating compute resources
- History of leading researchers through technical leaders while staying technically influential
- Proficiency in using AI tools as a core part of daily workflow and operations
- Ability to communicate complex technical tradeoffs to product and executive audiences
Nice to have
- Experience deploying TTS or generative-audio models at production scale
- Background in building or scaling high-performing AI research organizations
- Expertise in sophisticated evaluation systems for generative speech
- Recognized external contributions such as publications, patents, or open source
- Experience in fast-moving startup environments taking models from idea to production
Stack
- Text-to-Speech (TTS)
- Neural Audio Modeling
- Speech Generation
- AI Tools
How you'll work
- Fully remote position
- Fast-paced AI-first environment
- Requires adaptability and continuous learning