Audio Research Engineer (Remote — Worldwide)

Join the core team responsible for pushing the boundaries of generative music and high-fidelity audio synthesis at Suno. As an Audio Research Engineer, you will design and scale state-of-the-art transformer and diffusion models that handle complex multi-track musical structures and vocal timbre. This is a high-impact, remote role where your work will directly empower millions of creators to turn their ideas into broadcast-quality audio.

- Design and implement novel architectures for high-fidelity audio and music generation. - Stay at the forefront of AI research to integrate techniques like latent diffusion and autoregressive modeling into our production pipeline. - Develop efficient tokenization strategies for representating complex audio waveforms and MIDI-like structures. - Optimize model inference for real-time performance to support a global user base. - Collaborate with the product team to evaluate model quality through both objective metrics and subjective listening tests. - Build and maintain scalable data pipelines for processing terabytes of diverse audio datasets. - Contribute to internal research libraries and maintain high standards for code quality and reproducibility.

- PhD or Master’s degree in Computer Science, Machine Learning, or Audio Engineering with a focus on Deep Learning. - Proven track record of training large-scale generative models (Transformers, Diffusion, or VAEs) using PyTorch or JAX. - Deep expertise in Digital Signal Processing (DSP) and audio feature extraction (Spectrograms, Phase reconstruction, etc.). - Experience with large-scale distributed training across GPU clusters (H100s/A100s) and performance optimization. - Strong publication record at venues like NeurIPS, ICASSP, or ISMIR is a significant plus. - A passion for music and an understanding of music theory or production workflows. - Ability to work independently in a fully remote, globally distributed environment.

Equity, healthcare, music gear stipend.