Postdoc Position at Radboud University (Nijmegen) in NeuroAI: Biologically Grounded Neural Networks for Cognition

Can you help designing cognitive architectures to make brain implants smart in an international and interdisciplinary team involving the top neuroscience institutes in the Netherlands and thereby (ultimately) contribute to improve the life of patients?

Brain implants are, for the most part, fixed-function devices. The EBRAINS-NEUROTECH infrastructure is building a tool chain that changes this: neural networks that perform the perceptual and cognitive processing an implant needs, run on-device, and can be adapted by the experimenter.

You will build the core of that tool chain: a set of neural network modules for perception and cognition, grounded in what we know about the brain’s own mechanisms. For this position you will develop NeuroAI-in-practice: a set of neural network modules that can be easily used and adapted by an experimental user, while also giving publishable scientific insights in and of themselves. Two use-cases, involving processing for vision implants and implants for cognitive restoration, will guide the work. These networks also have to run on-device. Together with a postdoctoral researcher from the Centre for Mathematics and Computer Science (CWI), you will translate the designs to a neuromorphic architecture.

Does this sound like you?

• You are a scientific programmer or postdoctoral researcher with excellent programming skills and a strong interest in NeuroAI.
• You have a PhD degree in a quantitative field, such as physics, mathematics, computer science, engineering or computational or theoretical neuroscience.
• You have experience with programming in Python and C in combination with deep learning tools such as PyTorch, TensorFlow, and JAX.
• You have a background in neuroscience, preferably on mechanisms for cognitive processes, or a strong desire to obtain such a background. 

This position is physically hosted at the Neurophysics department of the Donders Institute at Radboud University. The work is part of the Software Platform of the EBRAINS-NEUROTECH infrastructure, involving teams at the University of Amsterdam (Dr Jorge Mejias), CWI (Prof. Sander Bohte) and Radboud University (Prof. Paul Tiesinga).

To apply for this job please consult the Academic Transfer link:

https://www.academictransfer.com/en/jobs/365079

For more information contact Paul Tiesinga at p.tiesinga@science.ru.nl