Research Scientist, Addressee Detection / Speech Separation

Role overview Advance research on identifying who speech is directed to and separating speech in crowded recordings where voices overlap. The work combines framing challenging modeling questions, running focused experiments, and translating findings into models suitable for deployment. The position is fully remote.

Responsibilities - Investigate addressee detection in complex, overlapping audio. - Develop and test approaches to speech separation in real-world recordings. - Frame research questions and design targeted experiments. - Turn promising results into models that can be deployed.

Requirements - The available description does not specify qualifications or experience requirements.