Projects
Certifications
Higher Diplomas
Languages
Industries
Let Me Introduce Myself
Data Scientist based in Paris, with a Master's in Data Science & AI from Centrale Lyon and an Engineering Diploma from ENSAM.
I build ML models that drive business decisions — from Marketing Mix Modeling and regression pipelines to deep learning for industrial imaging. I design automated dashboards, run statistical analyses that identify key performance levers, and deliver end-to-end data solutions across industries.
Education
2023 — 2024
Master in Data Science & Artificial Intelligence
Centrale School of Lyon · Ecully, France
2017 — 2023
State Engineer Diploma — AI & Data Science
National School of Arts and Crafts · Meknes, Morocco
Industrial Engineering applied to Artificial Intelligence & Data Science
Technical Skills
Experience
ROAS Optimization
Consistent campaign improvement
50+ Dashboards
Automated reporting pipeline
Deep Learning Engineer
Segmentation Accuracy
CT image classification
Products Shipped
Mobile + Web + Cloud
50% Faster
Page load optimization
Projects
Bank Customer Churn Prediction
ML model with Streamlit app for real-time churn predictions
Problem: Banks need to identify customers likely to churn to retain them proactively.
Approach: Trained multiple classifiers, deployed the best model as a Streamlit web app with real-time predictions.
Result: Integrated MLflow for experiment tracking and model versioning.
Fake News Detection Web Interface
Frequency analysis for real/fake news classification
Problem: Misinformation spreads rapidly online — automated detection is critical.
Approach: Built a voting classifier combining multiple NLP models with frequency analysis features.
Result: Improved classification accuracy by 5% over baseline.
Lane Departure Warning System
Real-time lane detection on Raspberry Pi using OpenCV and deep learning
Problem: Drivers need real-time warnings when unintentionally drifting out of lane.
Approach: Combined OpenCV edge detection with a lightweight deep learning model optimized for Raspberry Pi.
Result: Handles curves and varying lighting conditions in real-time on embedded hardware.
Image Similarity Web Interface
Django REST API for medical image comparison at 95% accuracy
Problem: Medical professionals need fast, accurate image comparison for diagnosis support.
Approach: Built a Django REST API with a classification model using K-means for feature extraction.
Result: 95% accuracy with 10% speed gain from K-means clustering optimization.
WebCam-Based Eye Tracking System
Real-time eye tracking with head movement compensation
Problem: Need for accessible eye tracking without specialized hardware.
Approach: Implemented real-time eye tracking algorithms with head movement compensation, exported tracking data to .dat files.
Result: Web-based system that tracks eye movement in real-time and saves results as images.
Flood Prediction Using Satellite Images
ML/DL classification of satellite imagery for flood area detection
Problem: Early flood detection is critical for disaster response and prevention.
Approach: Classified satellite images using deep learning with interactive map-based area selection.
Result: Over 75% accuracy with less than 5% error rate on flood area predictions.
Certifications
Applied Data Science with Python
University of Michigan
Applied AI with Deep Learning
IBM
TensorFlow Developer Specialization
DeepLearning.AI
Six Sigma Yellow Belt
6sigmastudy
Azure AI: Natural Language Processing
Microsoft
Azure AI: Computer Vision
Microsoft
Azure AI: Visual Tools for ML
Microsoft
Web Scraping with Scrapy & Python
Udemy
CSS Selector & XPath
Udemy
Python & Beautiful Soup with Flask
Udemy
Flutter & Dart [2022 Edition]
Udemy
Flutter Mobile AI Machine Learning
Udemy