Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : IRIMAS (Université de Haute-Alsace), MSD team, Mul
Durée : 6 months
Contact : maxime.devanne@uha.fr
Date limite de publication : 2027-01-08

Contexte :
Compact deep learning architectures and random-kernel methods such as MiniRocket reach state-of-the-art accuracy at a low computational cost. Their actual cost in time, memory and energy, as well as the reproducibility of their results, nevertheless depends on the software stack used, a point that is rarely studied in a controlled way.

Sujet :
To study how the programming language (Python, Julia, C) and numerical precision affect the cost and reliability of compact time series classifiers, without sacrificing accuracy: reproducing reference models, porting inference to other stacks, measuring cost and precision effects, statistical analysis, and open release of code and results.

Profil du candidat :
Master 2 or last year of Engineering School in Computer Science. Advanced skills in Python and C, machine learning libraries (scikit-learn, TensorFlow, PyTorch). Knowledge of Julia and of time series analysis will be appreciated.

Formation et compétences requises :
Master 2 or last year of Engineering School in Computer Science. Advanced skills in Python and C, machine learning libraries (scikit-learn, TensorFlow, PyTorch). Knowledge of Julia and of time series analysis will be appreciated.

Adresse d’emploi :
IRIMAS, Université de Haute-Alsace,
12 Rue des Frères Lumière
Mulhouse, France

Document attaché : 202611090952_internship_position_GL4TSC.pdf