Marouane Ait El Faqir

Machine Learning Engineer

Turning data into intelligent solutions

About

As a Machine Learning Engineer with specialized expertise in deep learning architectures, I solve complex problems through innovative machine learning solutions. My focus on self-supervised techniques delivers robust results in data-constrained environments.

Skills

Python
TensorFlow
PyTorch
Deep Learning
VAE GAN Jupyter Notebook Keras Scikit-learn Hyperparameter Tuning

Professional Experience

2020 - 2025

Data Scientist | AI & Machine Learning Specialist

CS Group, Toulouse

  • Developed deep learning and probabilistic models for space technology applications
  • Designed image super-resolution models for CNES (French Space Agency)
  • Built satellite maneuver prediction models for CNES and DGA (Defense Agency)
  • Contributed to Python libraries for ESA's Copernicus program
  • Implemented Variational Inference and Adversarial Generative Models for data analysis
2018 - 2020

Applied Mathematics Engineer

CHAUVIN-ARNOUX, Annecy

  • Designed numerical algorithms for embedded electronics using Python, C, and C++
  • Developed Bayesian deconvolution techniques for image filtering and 2D IIR filters
  • Built Gaussian process-based models for thermal camera image analysis
  • Led ML-based electromagnetic modeling for a portable current measurement device (patent filled)
2015 - 2018

Consulting Engineer | Machine Learning & Robotics

ALTRAN, Toulouse

  • Developed supervised learning models for turbulence prediction (Airbus R&D)
  • Designed ML-based drone control systems using deep learning
  • Applied Jacobian inverse kinematics for robotic control
  • Contributed to quality validation in aeronautics control laws engineering
2012 - 2015

PhD Researcher | Stochastic Optimization & AI

CNRS/INRA/Ecole Centrale de Lyon, Lyon-Paris

  • Developed stochastic optimization algorithms for large-scale biological systems
  • Applied Bayesian Optimization and Gradient-Based Methods for high-dimensional problems
  • Submitted findings in IEEE Conference On Decision and Control (CDC)
  • Created methods for proving optimization algorithms in large-scale systems biology

What I’m looking for

I am interested in roles and missions where advanced machine learning is applied to high-impact domains such as space, defense and embedded systems.

  • R&D or engineering positions focused on deep learning, probabilistic modelling and self-supervised learning.
  • Projects involving satellite imagery, orbital dynamics or critical systems monitoring.
  • Collaborations where machine learning bridges research ideas and operational constraints.

Application Domains

I apply machine learning and optimization techniques to demanding real-world environments such as space, defense, and embedded systems.

Space & Earth Observation

Super-resolution for Sentinel-2 optical imagery, satellite maneuver modelling, and contributions to Python libraries for the Copernicus (ESA/CNES) program.

Defense & Security

Predictive models for orbital dynamics, risk analysis and anomaly detection for critical systems in collaboration with DGA and national agencies.

Embedded Electronics & Sensors

Embedded algorithms for measurement instruments, probabilistic modelling for thermal cameras, and real-time signal processing.

Services

🤖

Deep Learning Architecture Design

Custom deep learning architectures including VAEs, GANs, and mixture models tailored to your specific needs.

📊

Anomaly Detection Systems

Development of robust anomaly detection solutions using state-of-the-art deep learning approaches.

💡

ML Model Optimization

Hyperparameter tuning and model optimization for improved performance and efficiency.

📚

ML Research & Development

Implementation of research papers and development of novel machine learning solutions.

Portfolio

Featured projects

L1BSR – Sentinel‑2 Super‑Resolution

Self-supervised super-resolution for Sentinel‑2 imagery, improving spatial resolution for downstream Earth observation tasks.

View Project

BiGAN Anomaly Detector

Bidirectional GAN-based anomaly detection system for complex, high-dimensional data distributions.

View Project

VAE Paper With Animation

Didactic implementation of Variational Autoencoders with animated visualizations of the latent space.

View Project

GMM-VAE

Gaussian Mixture Variational Autoencoder implementing a VAE variant with Gaussian Mixture as prior distribution.

Python Deep Learning VAE
View Project

VAE Paper With Animation

Interactive implementation of Variational Autoencoders with animated visualizations of latent space representations.

Python Deep Learning VAE Animation
View Project

Expectation-Maximization

Interactive explanation of the Expectation-Maximization algorithm with application on MNIST dataset classification.

Python Machine Learning EM Algorithm MNIST
View Project

BiGAN Anomaly Detector

Advanced anomaly detection system based on Bidirectional Generative Adversarial Network architecture.

Python GAN Deep Learning
View Project

L1BSR - Sentinel-2 Super-Resolution

Self-supervised single-image super-resolution pipeline for Sentinel-2 L1B optical imagery.

  • Designed a fully self-supervised training strategy without high-resolution ground truth.
  • Improved spatial resolution for tracking and detection tasks in Earth observation.
  • Integrated into an operational satellite imagery processing workflow.
Python Computer Vision Super-Resolution
View Project

VAE-GAN

Implementation of VAE-GAN architecture based on the paper "Autoencoding beyond pixels using a learned similarity metric".

Python VAE GAN Deep Learning
View Project

Neural E-M

Implementation of "Neural Expectation-Maximization" for unsupervised learning of latent variables.

Python Deep Learning Unsupervised Learning
View Project

Keras Tuner Works

Training discriminative models on time series data with hyperparameter tuning using KerasTuner library.

Python Keras Time Series Hyperparameter Tuning
View Project

Contact

For any opportunity, consulting mission or collaboration, feel free to contact me directly.

Send an email View my LinkedIn profile