Two open postdoc positions.

Postdoc · Role 1

Diffusion world models for psychiatry

What you'd work on

Building diffusion world models for psychiatry: generative models that capture the joint dynamics of brain activity, facial video, and language over a course of treatment, and use them to reason about treatment counterfactuals.

Qualifications

• PhD in CS, ML, applied math, or a related quantitative field.
• Experience with generative or sequence models (diffusion, world models, transformers) and large multimodal datasets.
• Strong Python and modern deep-learning frameworks (PyTorch).
• Genuine interest in clinical psychiatric questions, not just methods.

Postdoc · Role 2

Interventional generative twins for neurostimulation

What you'd work on

Building interventional generative twins for neurostimulation: patient-specific models that predict how an individual's brain responds to stimulation, and turn those predictions into personalized, closed-loop TMS.

Qualifications

• PhD in computational neuroscience, ML, biomedical engineering, or a related quantitative field.
• Experience with neural or clinical data (high-density EEG, fMRI, or electrophysiology); TMS or closed-loop stimulation a plus.
• Ideally, an interest in neural operator design for inverse problems.
• Strong Python and PyTorch, and an interest in translating models into clinical intervention.