Researcher in Artificial Intelligence / Machine Learning - Lead Data Scientist.
Education
Habilitation in Computer Science 2017
\rightarrow Chemnitz University of Technology, Germany.
Ph.D. in Computer Science 2006
\rightarrow Université Henri Poincaré Nancy-I, France.
Engineering degree in Signal Processing and Microelectronics 2002
\rightarrow École Supérieure d’Électricité (Supélec), Rennes, France.
Master of Science in Microelectronics 2002
\rightarrow Université Rennes-I, France.
Experience
eOdyn - France 2025 - now
\rightarrow AI researcher and lead Data Scientist at eOdyn in Plouzané / Brest.
- Nereus: short-term forecasting of ocean surface currents with generative deep learning (flow matching, U-Net, Swin Transformer, Fourier Neural Operators).
- OSmose: data assimilation fitting AIS observations onto a physics-constrained multiscale generative model and an altimetric background.
- Anomaly detection in AIS signals.
- EU project AI4COPSEC.
Chemnitz University of Technology - Germany 2011 - 2025
\rightarrow Researcher / lecturer in the faculty of Computer science, AI lab of Fred Hamker.
- Research:
- Neuroinformatics: neurosimulator ANNarchy.
- Deep reinforcement learning.
- Reservoir computing.
- Computational neuroscience: basal ganglia, dopamine, hippocampus.
- Teaching:
- Neurocomputing / Machine Learning
- Deep Reinforcement Learning
- Computer Vision
- Introduction to AI.
- Erasmus+ departmental coordinator.
WWU University of Münster - Germany 2006 - 2011
\rightarrow Postdoc in the Institute of Psychology, lab of Markus Lappe, supervisor Fred Hamker.
- Research:
- Computational neuroscience: basal ganglia.
Inria Lorraine - France 2002 - 2006
\rightarrow PhD student in the Cortex team (LORIA Nancy), supervisors Frédéric Alexandre and Nicolas Rougier (now in Mnemosyne).
- Research:
- Computational neuroscience: basal ganglia, visual attention.
- Robotics.
- Teaching (at the University of Nancy-I):
- Computer architecture, Java, AI.
Skills
Programming languages: Python, C++, C, Java, Matlab, Julia
Machine learning frameworks: pytorch, tensorflow / keras, scikit-learn, pandas, numpy, XGBoost, rllib, tianshou
Deep learning: transformers, U-Net, Swin Transformer, Fourier Neural Operators, RNN / LSTM, deep RL
Generative AI: flow matching, diffusion models, ollama
Parallel computing: OpenMP, MPI, CUDA, SLURM clusters
Tooling: linux, git, docker, mlflow, wandb
Third-party projects
Title: Security enhancement through heterogeneous data fusion and improved AI/ML-powered Copernicus maritime and border surveillance services.
Partners: Simula Research Laboratory (coordinator), eOdyn (technical coordinator), IMT Atlantique, SatCen, Università Politecnica delle Marche, Guardia Civil, Hozint, ITML, Aistech Space, Magellan Circle.
Role: Design of the surface current forecasting and data assimilation components.
Title: Counterfactual Assessment and Valuation for Awareness Architecture.
Partners: Radboud University, TU Chemnitz, CERTH, Sorbonne Université, Eodyne Systems, University of Sheffield, Robotnik Automation, University of Oxford, Uppsala University, tp21 GmbH.
Role: Research.
Title: Erforschung von KI Methoden zur Entwicklung eines interaktiven Gesundheitsassistenten auf Basis von Human in the Loop Machine Learning.
→ Research on AI methods for the development of an interactive health assistant based on human-in-the-loop machine learning.
Partners: Corant GmbH, TU Chemnitz.
Role: Design, co-supervision of Aida Farahani and Payam Atoofi.
Title: Wissensbasierte Anomalieerkennung mittels Künstlicher Intelligenz in Kritischen Infrastrukturen.
→ Knowledge-based anomaly detection using artificial intelligence in critical infrastructures.
Partners: TU Cottbus, TU Chemnitz, LEAG, RWE AG, STEAG GmbH, ASCORI GmbH, migosens GmbH, ZEDAS GmbH.
Role: Design, co-supervision of René Larisch.
Title: Maschinelles Lernen zur Prognose von Prozessparametern und Bauteilqualität in der automobilen Karosserieproduktion.
→ Machine learning for the prediction of process parameters and component quality in automotive body production.
Partners: Fraunhofer IWU Dresden, Scale GmbH, TU Chemnitz.
Role: Design, research, co-supervision of Aida Farahani and Payam Atoofi.
Title: Auto-tuning for neural simulations on different parallel hardware.
Partners: TU Chemnitz (Hamker).
Role: Design, co-supervision of Helge Dinkelbach.
Title: Sozial agierende, kognitive Systeme zur Feststellung von Hilfsbedürftigkeit.
→ Social, active and cognitive systems for the detection of helplessness
Partners: TU Chemnitz.
Role: Principal Investigator
Title: StayCentered - Methodenbasis eines Assistenzsystems für Centerlotsen - MACeLot.
→ Method base of an assistance system for center pilots
Partners: TU Chemnitz.
Role: Research.
Title: CrossWorlds: Kopplung virtueller und realer sozialer Welten
→ CrossWorlds: Coupling virtual and real social worlds.
Partners: TU Chemnitz.
Role: Principal investigator.
Title: The cognitive control of visual perception and action selection.
Partners: TU Chemnitz (Hamker).
Role: Research.
Title: A systems approach to modeling the cognitive guidance of attention and object/category recognition.
Partners: TU Chemnitz (Hamker).
Role: Research (employed).
Title: Biomimetic Multimodal Learning in a Mirror Neuron-based Robot.
Partners: Sunderland, Inria, Parma, Cambridge, Ulm.
Role: Research (employed).