Postdoctoral Position in AI, Machine Learning and Auditory Neuroscience

Duration: 24 months (flexible start)

Supervisor: Keith Doelling (INSERM)

About the position

We are recruiting a postdoctoral researcher to focused on exploration and development of AI models of auditory perception, towards a broader goal of understanding how the brain predicts and learns from human communication sounds such as speech and music. The project sits at the intersection of machine learning, neuroscience, and behavior, with applications ranging from music cognition to speech-in-noise perception in audiology. You will work with naturalistic audio, large datasets, and multimodal neural recordings (EEG, fMRI, intracranial) to develop models with a major goal of interpretability so that we may better understand the influence of complex, natural communication signals on our mental processes.

What you will do

• Build novel deep learning architectures for auditory prediction (speech and/or music) prioritizing explainability and cognitive hierarchies
• Quantify prediction, uncertainty, and learning in real-world stimuli and track how they are encoded in neural data
• Develop tools to be used in audiological clinical to diagnose patients based on naturalistic interactions.
• Work with an interdisciplinary team to better understand neural function in healthy and patient populations.
• Communicate your work with non-experts in ML domains to facilitate collaboration
• Publish peer-reviewed articles based on your findings.

You will have flexibility to shape your research direction within the broader theme of predictive processing.

Why join this project?

This position offers something different from standard ML roles. Rather than focusing only on performance, we aim to understand how models work and what they reveal about the brain.

You will:

• Own your project and drive ideas from concept to publication
• Work on problems with scientific depth and long-term impact
• Develop models that influence both neuroscience and AI research

This is an ideal position if you want to:

• build an independent academic profile, or
• strengthen your expertise for roles in industry or startups

The combination of technical depth, interdisciplinary research, and intellectual ownership provides a strong platform for future career development in both academia and tech.

Profile Required:

• PhD in ML, computational neuroscience, physics, engineering, or related field
• Strong experience in machine learning (PyTorch or equivalent)
• Experience with sequence or time-series models

Bonus:

• Audio, speech, or music processing
• Transformers, reinforcement learning, or probabilistic models
• Interest in brain and behavior

You will join the Human and Artificial Perception (HArP) group within the Institut de l’Audition, a center of Institut Pasteur focused on sound processing and hearing health. HArP is a new team focused on real-world auditory perception, aiming to better understand how the brain processes sound outside of the lab and developing diagnostics tests for perception outside of the clinic. We are a new but highly motivated and vibrant team of experimentalists and computationalists devoted to leveraging state-of-the-art tools to develop novel insights into how humans hear

We offer:

• Access to neuroimaging, clinical collaborations, and large datasets
• Strong connections between AI and neuroscience communities
• The Pasteur Institute is committed to providing an inclusive and respectful work environment. We welcome all applications that meet the job requirements, valuing diversity in all its forms and a variety of career paths.
• Very favorable vacation and RTT policy (204-day package)
• Sustainable transportation allowance or 75% reimbursement of transit passes
• Family health insurance (free for spouses and children) and on-campus health services
• Strong CSR commitments, and employee initiatives encouraged
• Multicultural environment (more than 75 nationalities on campus)

Apply:

Please send your CV, cover letter (1-2 pages), 2–3 references, to Keith Doelling ( )