Portfolio

Research &
industrial projects.

Complex problems, rigorous methods, concrete results.

2026 — present
project_04
Eliis

R&D engineer
at Eliis

Developing new features for PaleoScan, Eliis's seismic-interpretation platform, to improve the analysis of geological images — deep learning and image processing applied to seismic data and stratigraphy.

  • Designing and building an end-to-end research project
  • Seismic facies analysis with deep learning
  • Leveraging seismic attributes derived from RGT (Relative Geological Time)
  • Generating facies data with diffusion models
PaleoScan Deep Learning Diffusion models Seismic facies RGT
Learn more Eliis · PaleoScan ↗
2021–2025
project_03
DGA · Thales · CNRS

Rare-object detection
in sonar imagery

Development of object-detection models for low-sample sonar images, combining deep neural networks with smart augmentation strategies.

  • YOLO models adapted to low-resolution sonar images
  • GANs for synthetic data generation
  • Active learning combining uncertainty and clustering to reduce annotation needs
  • Significant detection-accuracy gains with less data
PyTorch YOLO GANs Active learning Python
2019–2021
project_02
ANR · DGA ASTRID

Image steganalysis
with neural networks

As part of the ANR DGA ASTRID project, development of deep-learning methods to detect hidden messages in large-scale JPEG images.

  • Development in PyTorch and TensorFlow
  • Handling massive JPEG image databases
  • Creation of the LSSD database (Large JPEG Steganalysis Dataset)
  • Publication: “LSSD: A Controlled Large JPEG Image Database for Deep-Learning-Based Steganalysis” — ICPR 2021
PyTorch TensorFlow Steganalysis JPEG Deep Learning
Related publications LSSD (ICPR 2021) Scalability (ICPR 2021) CORESA 2020 LSSD database →
2018
project_01
Defence · Embedded systems

Automatic guidance
of hypersonic missiles

Applying reinforcement learning to design agents able to autonomously guide hypersonic missiles in simulated environments.

  • Implementation of DDPG (Deep Deterministic Policy Gradient)
  • Implementation and comparison with PPO (Proximal Policy Optimization)
  • Simulation and validation in Python and MATLAB
  • Flight-policy optimisation under constraints
Reinforcement Learning DDPG PPO Python MATLAB

Availability

Have a project in mind?

I'm available for collaborations, contracts or applied-research positions.

Get in touch →