Multimodal AI Researcher, Audio

You will push the boundaries of the state-of-the-art in audio and multimodal technologies. The ideal candidate would have a strong background in deep learning, both in terms of conceptual understanding, as well as practical experience, with previous exposure to audio applications. A core aspect of this role involves being able to keep up to date with the literature, implement, and innovate with the bleeding edge in generative models, self-supervised learning, and multi-modal learning.

With the explosion of large language models and natural language processing, you will partner closely with Dolby's worldwide AI research staff, which actively pursues the integration of such models into audio and media experiences. You will be able to hit the ground running, innovate, and contribute to such projects. Consequently, experience with language models, question answering, vision-language models, captioning, etc. would be highly beneficial.

We are looking for candidates with experience in any of the following:

• Generative modeling for audio applications (diffusion models, autoregressive models, masked generative transformers).
• Multimodal semantic understanding and multimodal reasoning.
• Multimodal representations (audio-video, audio-text, audio-video-text).
• Multimodal AI architectures, with a focus on generating audio, music, and speech (text-to-audio, video-to-audio, image-to-audio).
• Self and semi-supervised learning.
• AI driven audio enhancement, processing, and generation (for speech and music), such as speech enhancement and analysis, source separation, text-to-speech, text-to-music, music information retrieval, audio classification.
• LLMs for audio applications.

• Partner closely with other domain experts to refine and execute Dolby's technical strategy in artificial intelligence and machine learning.
• Use deep learning to create new solutions (including foundation models) and enhance existing applications.
• Push the state-of-the-art and develop intellectual property.
• Transfer technology to product groups.
• Establish research collaborations with external university partners.
• Mentor interns on novel research problems.
• Publish papers in top-tier conferences and journals.
• Advise internal leaders on recent deep learning advancements in the industry and academia to further influence research direction and business decisions.

• Ph.D. in Computer Science or similar field.
• A strong background in deep learning, both in terms of conceptual understanding, as well as practical experience.
• Technical knowledge of audio fundamentals.
• Deep passion for audio, music, and multimedia applications.
• Deep knowledge on current machine learning literature.
• Strong publication record, with publications in major machine learning conferences (e.g. NeurIPS, ICLR, ICML) or top domain-specific conferences is desirable (e.g., ACL, CVPR, ICASSP, Interspeech).
• Highly skilled in Python and one or more popular deep learning frameworks (TensorFlow or PyTorch).
• Ability to envision new technologies and turn them into innovative products.
• Good communication and collaboration skills.