🕒 9 days ago
Design and validate rigorous, multi-step computational life sciences tasks that test whether AI agents can inspect scientific data, write and run code, troubleshoot workflows, and produce verifiable results.
Research Scientist - Generative Audio
📅 Jul 10
Generative AudioDiffusion ModelsFlow MatchingVocal Synthesis
📜 Description
- Research generative audio using diffusion or flow-matching models, focusing on vocal synthesis, post-training, or music editing.
- Develop preference-alignment methods, reward models, and reinforcement learning approaches to improve music-generation quality and controllability.
- Explore iterative music generation and audio editing, including stem, instrumentation, mood, tempo, and structure changes.
- Run large-scale experiments and create practical applications for new listening experiences.
- Collaborate with scientists, engineers, product managers, designers, user researchers, and analysts.
- Publish findings, deliver talks, and participate in leading research conferences.
🛠️ Requirements
- Ph.D. in Computer Science, Mathematics, Engineering, or a related field; industry experience is helpful.
- Experience in generative modeling, machine learning, music information retrieval, speech or audio processing, signal processing, probabilistic modeling, computer vision, or a related field.
- Deep expertise in vocal or speech synthesis, post-training alignment, or audio-to-audio generation and text-guided music editing.
- Publications at leading conferences such as ICASSP, ISMIR, INTERSPEECH, ICLR, AAAI, IJCAI, NeurIPS, ICML, CVPR, ECCV, or ICCV.
- Strong coding skills in Python, PyTorch, and NumPy.
- Ability to explain complex topics clearly, build strong working relationships, and translate research ideas into products.
Full job description
We are seeking Research Scientists (across all levels of seniority) to join our Artist-First AI Music Lab. Our team pioneers and advances state-of-the-art generative technologies for music that create breakthrough experiences for fans and artists. We invent entirely new listening experiences that center and celebrate artists and creatives. All of our products will put artists and songwriters first, through these four principles:
- Partnerships with record labels, distributors, and music publishers: We’ll develop new products for artists and fans through upfront agreements, not by asking for forgiveness later.
- Choice in participation: We recognize there’s a wide range of views on the use of generative music tools within the artistic community. Therefore, artists and rightsholders will choose if and how to participate to ensure the use of AI tools aligns with the values of the people behind the music.
- Fair compensation and new revenue: We will build products that create wholly new revenue streams for rightsholders, artists, and songwriters, ensuring they are properly compensated for the use of their work and transparently credited for their contributions.
- Artist-fan connection: AI tools we develop will not replace human artistry. They will give artists new ways to be creative and connect with fans. We will leverage our role as the place where more than 700 million people already come to listen to music every month to ensure that generative AI deepens artist-fan connections.
Learn more in our press release: https://newsroom.spotify.com/2025-10-16/artist-first-ai-music-spotify-collaboration/
What You'll Do
Conduct groundbreaking research in generative audio using diffusion or flow matching models, with a focus on one or more of the following areas:
Vocal Synthesis — Research in vocal and speech synthesis, along with related areas such as ML-based audio processing and signal processing.
Post-Training — Research in post-training techniques for music generation, including preference alignment methods (such as DPO, RLHF, or KTO), reward model design and training, and reinforcement learning to improve output quality, controllability, and human preference adherence.
Editing — Research in iterative music generation and audio editing, including capabilities such as stem replacement, instrumentation change, mood changes (while preserving content), tempo changes, and structure changes.
You will also:
Who You Are
Where You'll Be
Similar jobs
Search more Research Scientist jobs
🕒 22 days ago
Develop and evaluate machine learning models across the full research lifecycle, from novel architectures and training methods to evaluation protocols and efficient training and inference, while building research code for rapid.
Quantitative Researcher (Mulitple Openings)
🕒 12 days ago
Tower Research Capital🏢 Financial Services🤝 B2B
🕒 12 days ago
Research and deploy trading algorithms by analyzing market data, developing statistical signals and tools, and improving simulations to support high- to mid-frequency trading and live execution.
Epidemiologist, Real-World Evidence (Remote US)
🕒 4 days ago
International-Consulting-Associates-Inc👥 11 - 50 employees🏢 Management Consulting
🕒 4 days ago
Lead client and stakeholder collaborations to assess real-world data sources and generate robust evidence for medical-device regulatory decisions, drawing on expertise in healthcare data, epidemiology, and RWD analytics.
🕒 5 days ago
Lead disability employment research, designing rigorous studies, analyzing data, and engaging people with disabilities and community organizations to inform employment supports and outcomes.