Research Scientist - Generative Audio

Spotify
London, Stockholm
Workplace: RemoteFull timeFunction: Data Science & Machine LearningEducation: phdSkills: ["Creative problem-solving","Communication","Collaboration","Stakeholder management"]

Conduct cutting-edge research in generative audio for Spotify’s Artist-First AI Music Lab, using approaches such as diffusion or flow matching. Focus on vocal/speech synthesis, post-training alignment (e.g., DPO, RLHF, PPO/GRPO/KTO), and iterative audio editing such as stem replacement and tempo/mood/structure changes. Run large-scale experiments, build practical applications at scale, collaborate across scientists and engineers, and publish findings at top conferences.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Spotify
Spotify
1 month ago

Research Scientist - Generative Audio

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 7 hours agoStatus: Live

Job Summary

Conduct cutting-edge research in generative audio for Spotify’s Artist-First AI Music Lab, using approaches such as diffusion or flow matching. Focus on vocal/speech synthesis, post-training alignment (e.g., DPO, RLHF, PPO/GRPO/KTO), and iterative audio editing such as stem replacement and tempo/mood/structure changes. Run large-scale experiments, build practical applications at scale, collaborate across scientists and engineers, and publish findings at top conferences.
Location: London, Stockholm
Workplace: Remote
Employment Type: Full time
Job Function: Data Science & Machine Learning

Key Responsibilities

  • •Conduct generative-audio research using diffusion or flow-matching approaches across vocal/speech synthesis, post-training, and iterative editing.
  • •Develop post-training techniques for music generation, including preference alignment, reward model design/training, and reinforcement learning to improve output quality and controllability.
  • •Research and implement iterative music generation and audio editing capabilities such as stem replacement, instrumentation/mood/tempo/structure changes.
  • •Run large-scale experiments using Spotify’s infrastructure and build practical applications that advance listening experiences.
  • •Collaborate cross-functionally to craft innovative solutions and share results through publications, talks, and conferences.

Key Requirements

  • •Hold a Ph.D. in Computer Science, Mathematics, Engineering, or a related field.
  • •Experience in generative modeling and/or machine learning, including areas such as music information retrieval, speech/audio processing, signal processing, probabilistic modeling, or computer vision.
  • •Deep expertise in one or more focus areas including vocal/speech synthesis, post-training alignment (e.g., PPO, GRPO, DPO), or audio-to-audio generation and text-guided music editing.
  • •Strong record of publications at leading conferences such as ICASSP, ISMIR, INTERSPEECH, ICLR, NeurIPS, ICML, CVPR, or ECCV.
  • •Strong coding skills in Python, PyTorch, and NumPy.
Experience:Machine learningGenerative modelingMusic information retrievalSpeech processingAudio processingSignal processingProbabilistic modelingComputer vision
Education:PhD / Doctorate
Skills:Creative problem-solvingCommunicationCollaborationStakeholder management
Tech Stack:Diffusion modelsFlow matchingGenerative modelingMachine learningSpeech processingAudio processingSignal processingProbabilistic modelingComputer visionReinforcement learningPPOGRPODPORLHFKTOText-guided music editingPythonPyTorchNumPyICASSP

Company Brief

Spotify
Provides a music, podcast and audio streaming service offering on-demand access to millions of tracks and podcasts, personalized recommendations, and advertising and subscription-based tiers for listeners and creators worldwide.
Industry: Streaming Platforms
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Public Company
Valuation: Public Company (Market Cap in USD)
Funding: IPO / Publicly Listed
Headquarters: Stockholm, Sweden
Founded: 2006
Glassdoor
Glassdoor: 4.1
WebsiteLinkedInGlassdoor