Uberduck makes AI vocal generation tools for music professionals, hobbyists, and marketers.

Vocal samples and full songs produced by our text-to-speech, text-to-singing, text-to-rapping, and voice conversion technologies power:

• AI remixes and original songs for musicians ranging from hobbyists to the largest DJs in the world.
• Personalized musical content in marketing campaigns.
• A variety of music-related consumer experiences built on top of our APIs.

Uberduck is seeking an ML research engineer to help us develop AI software for singing, rapping, and music synthesis. You'll help scale our ML stack—algorithms, data, and compute—and rapidly deploy new research to an audience of millions.

The ideal candidate has at least two years experience working in audio machine learning, experience deploying deep learning models to production, and a love for programming and music.

In person at our Seattle office strongly preferred, but not a dealbreaker.

Frontend: Typescript, React, NextJS Backend: Python, FastAPI, Postgres, Redis, Kubernetes ML: PyTorch, Ray

Things you might work on:

• ML inference pipelines that serve thousands of models with low latency.
• Zero shot voice conversion and zero shot text-to-singing voices.
• User-friendly interfaces for searching across thousands of voices.
• ML-first audio tooling.