PhD: Pretrained Multimodal Neuroimaging for Cognition

1000scholars; Palaiseau; On-site; EUR 16,000 - 23,000; Full time

Job summary

1000scholars seeks a PhD candidate to join the MIND Inria team at CEA NeuroSpin on the Paris-Saclay campus to develop pretrained multimodal neuroimaging models with calibrated uncertainty.

The PhD will build data infrastructure for large datasets (HCP, UK Biobank), design deep generative and self-supervised models, and compare transfer learning strategies across datasets, with results shared openly through benchmarks and released code.

Qualifications

• PhD in neuroscience, computer science or related field with strong ML background.
• Experience with neuroimaging data and Python ML stack is desirable.
• Ability to collaborate with international teams and present at conferences.

Responsibilities

• Build and release open preprocessing and harmonisation pipelines for large multimodal neuroimaging databases.
• Design, train and benchmark deep generative and self-supervised models of the multimodal brain phenotype.
• Evaluate transfer learning strategies on small to medium datasets and release open benchmarks.
• Validate research results and publish at leading conferences and journals.
• Release models and code openly and collaborate with partner teams.

Skills

Python programming

Deep learning (PyTorch/JAX)

ML theory (probability / statistics)

Neuroimaging data analysis

Data pipelines

Education

PhD in Neuroscience / ML / AI

Tools

PyTorch

JAX

MRtrix

DIPY

Nilearn

Job description

1000scholars seeks a PhD candidate to join the MIND Inria team at CEA NeuroSpin on the Paris-Saclay campus to develop pretrained multimodal neuroimaging models with calibrated uncertainty.

The PhD will build data infrastructure for large datasets (HCP, UK Biobank), design deep generative and self-supervised models, and compare transfer learning strategies across datasets, with results shared openly through benchmarks and released code.