Research Scientist - Generative Audio

📋 Description

• Conduct groundbreaking research in generative audio using diffusion or flow matching.
• Focus areas include vocal synthesis and ML-based audio processing.
• Research post-training alignment techniques (DPO/RLHF/KTO) and reward models.
• Research audio editing and iterative music generation (stem/mood/tempo edits).
• Run large-scale experiments with Spotify's infrastructure and 700M MAU.
• Collaborate with cross-functional teams to craft innovative listening solutions.

🎯 Requirements

• PhD in CS, Mathematics, Engineering, or related field.
• Experience in generative modeling, ML, MIR, or audio/speech processing.
• Deep expertise in vocal synthesis, post-training, or audio-to-audio editing.
• Publications at top venues (ICASSP, INTERSPEECH, NeurIPS, ICML).
• Strong coding in Python, PyTorch, NumPy.
• Creative problem solver who turns research into scale products.

🎁 Benefits

• Flexible remote work within Europe with CET/EST collaboration.
• Access to Spotify’s infrastructure and 700M MAU audience.
• Opportunities to publish findings at top conferences.
• Equal opportunity employer; accommodations available.