About the Role

Work at the centre of how Mirelo AI builds its next generation of multimodal video-to-audio models, exploring and building new multimodal models to push the boundaries of music, sound and speech generation.

Key Responsibilities

Design, implement and train large-scale multimodal generative models for audio generation (diffusion and/or autoregressive)
Explore new modeling ideas for audio generation, drawing on language and image domains
Develop and experiment with post-training for new capabilities (fine-grained control, in/out-painting, editing)
Conduct rigorous ablation studies and communicate results to the team
Contribute to all stages of model development: data curation, experimentation, evaluation and deployment

Requirements

Hands-on experience training large-scale generative models in a fast-paced research environment
Deep understanding of cutting-edge ML research in image, language, video or audio
Strong proficiency in PyTorch, transformer architectures and modern deep learning tooling
Solid understanding of distributed training techniques (FSDP, low precision training, model parallelism)
Strong track record in generative models (publications, open-source, or applied ML projects)

Details

Hybrid role based in Berlin or Tübingen, Germany. Mirelo AI recently closed a $41M Seed round co-led by Andreessen Horowitz and Index Ventures.

Research Scientist - Model Team