Member of Technical Staff - Quantitative Research

Udio
New York City (Remote possible for exceptional candidates) / New York, New York, New York, United States2026-01-22

About the job

We are looking for a full-stack scientist to pioneer quantitative research efforts at Udio. You will build at the intersection of research, engineering and product, bridging disciplines by drawing on huge, one-of-a-kind proprietary datasets of music, metadata and user interactions/feedback. Working closely with the modeling team, product leadership and the music evaluation manager, you will apply your research toward pushing the frontier of music generation, setting a course through a bleeding-edge product category and unlocking new revenues for artists and experiences for fans.

Responsibilities

Design & own evaluation/optimization frameworks for frontier music models

Drive product & research roadmap

Build stable infrastructure

Champion scientific rigor

Qualifications

Minimum

Deep quantitative expertise: Ph.D. in statistics, mathematics, physics, or another quantitative discipline, or 5+ years’ industry experience as a quantitative analyst / data scientist

Autonomy & ownership: You thrive in greenfield research domains, undefined product categories and small, flat teams. Driven by curiosity and good taste, you ask good questions in addition to finding good answers.

Engineering chops: You’re adept in translating your ideas into clear, production-ready code and collaborating in an active research codebase.

Excellence in scientific communication: You relate technical information with rigor and crystal clarity to researchers, engineers, product managers and business partners alike.

Preferred

Obsession with music & the science of sound. Experience in DSP, MIR, music production / composition / performance, and a big record collection all a huge plus.

Familiarity with deep learning frameworks, especially JAX.

Experience with GCP, Apache Beam/DataFlow, Kubernetes, TensorFlow Data / TFRecord.

Experience designing evaluation frameworks specifically for generative model outputs in any modality.