Working student (all genders) - Interpretable Machine Learning for EEG-based Classification

Augsburg, Berlin, Erlangen, Ingolstadt, Karlsruhe, Krumbach, Leipzig...

Tiempo completo o tiempo parcial

Estudiante en prácticas

Descripción corta

In the EXACT-EEG research project, we are developing an interpretable machine learning system for automatic classification of electroencephalography (EEG) data. Our goal is to facilitate and accelerate neurological diagnostics through interpretable AI methods and to make patterns visible that have previously been undetectable using conventional approaches. The system builds on a Self-Explaining Selective Model (SESM) architecture that learns compact, class-specific concepts from raw EEG signals. These concepts serve directly as explanations for each prediction, showing which time-domain and frequency-domain features were decisive for the classification. At XITASO, we are advancing this research by

designing and evaluating novel model variants that extend the dual-stream time-frequency SESM architecture,
running systematic architecture evaluations on a SLURM cluster,
analysing experiments and translating them into reproducible evaluation protocols, and
improving the explainability of model decisions through enhanced attention-based concept representations.

Estas tareas te interesan

Further development of existing model variants and training pipelines in the EXACT-EEG codebase.
Implementation and evaluation of new architectural ideas for interpretable EEG classification (e.g. multi-channel analysis, cross-montage generalization).
Systematic experimentation on a SLURM-managed GPU cluster and analysis of results across architecture and hyperparameter grid searches.
Development of analysis tools, visualisations, and notebooks to evaluate concept quality and model interpretability.

Estas son tus cualidades

You are completing a degree in (medical) computer science, data science, mathematics, or a comparable subject at a university.
You have experience with Python and PyTorch. Familiarity with PyTorch Lightning, Hydra, MNE, or SLURM is a plus but not required.
You have a basic understanding of machine learning concepts; prior exposure to deep learning for time-series, explainable AI, or biomedical signals is welcome.
You are curious, want to work in a technologically advanced environment and are looking for an opportunity to apply and deepen the knowledge you have gained during your studies.
You enjoy open and honest communication with colleagues and research partners, offer constructive criticism and accept feedback.
Your language skills enable you to perform your role in English (at least C1-level). German is desirable but not essential.

Información salarial

Within our standardized and transparent salary framework, the pay for this position ranges from €15.50 to €19.50 per hour and is based on various factors, such as qualifications and experience.

Tu persona de contacto

Daniela
+49 821 885882-0
work@xitaso.com