Machine Learning TTS Engineer
Core
Building next-generation LLM-based Text-to-Speech (TTS) systems for new languages and moving research models to production quality.
Role type
Senior Machine Learning TTS Engineer
Builds
Production-ready TTS models and audio data pipelines for voice assistants and devices
Domain
Speech Technology, Generative AI, NLP
Deliverable
production ML models
Required skills
LLM-based TTS (VALL-E, CosyVoice, XTTS), PyTorch, deep learning (Transformers, diffusion/flow matching), neural audio codecs, vocoders, streaming TTS architecture, audio signal processing, text normalization, speech data processing, TTS evaluation (MOS, SBS), Python
Preferred skills
Low-resource language experience, dataset creation, MLOps (MLflow, DVC), C++, publications
Technologies
PyTorch, MLflow, DVC, VALL-E, CosyVoice, XTTS
Responsibilities
Extend TTS to new languages via data collection, curation, and model adaptation; take TTS architectures from research to production; push models via post-training and architectural improvements; build efficient audio data pipelines; design rigorous benchmarks and evaluations
Seniority
Senior, hands-on IC
