Summary

PhD student in Artificial Intelligence at Université Paris-Saclay (Lincoln lab). Research focused on real-time waste detection via drone using federated learning and multi-objective optimization. Also teaching Mathematics, AI, and MLOps at École Centrale d'Electronique.

Education

• Mathematics and Computer Science

  2027
  Université Paris-Saclay
• Mathematics, Vision, Learning (MVA)
  ENS Paris-Saclay

  Courses: Large Language Models, Detection Theory and Industrial Applications, Convex Optimization and Applications in Machine Learning, Foundations of Distributed and Large-Scale Computing Optimization
• Mathematics - Computer Science - Data Science (MIDS)

  2023
  Université Paris Cité

  GPA: >14/20
• Mathematics

  2021
  Université Paris Est Créteil
• Speciality Mathematics

  2018
  Lycée Emily Brontë - Lognes

Work Experience

• PhD Student in Artificial Intelligence

  2024-04-01 -
  Lincoln / Université Paris-Saclay

  Research on real-time and continuous waste detection by drone via federated learning and multi-objective optimization. MVA Master at ENS Paris-Saclay (LLMs, Detection Theory, Convex Optimization, Distributed Computing).

  • Real-time waste detection by UAV using federated learning and multi-objective optimization
  • MVA courses: Large Language Models, Detection Theory, Convex Optimization, Distributed Computing
• Professor

  2025-09-01 -
  École Centrale d'Electronique

  Teaching Mathematics (Prépa, bac+1), Artificial Intelligence (Engineering cycle, bac+5), and MLOps (MSc AI, bac+5).

  • Mathematics: Preparatory class (bac+1)
  • Artificial Intelligence: Engineering cycle (bac+5)
  • MLOps: MSc AI (bac+5)
• Data Core Team

  2024-03-01 - 2025-09-01
  AtlasAI

  Data collection and preparation for training language models for the Moroccan dialect. LLM training for Moroccan dialect (data and model available at atlasai.com).

  • Data retrieval and preparation for Moroccan dialect LLM training
  • LLM fine-tuning with Hugging Face
• Deep Learning Engineer

  2023-10-01 - 2024-04-01
  Lincoln

  Generative AI: Multi-Modal Retrieval Augmented Generation (RAG) — patented. Document structuring by AI. CI/CD deployments on Azure DevOps.

  • Multi-Modal RAG pipeline (LLM, embedding model, vector database, inference engine) — patented
  • Synthetic question/answer data generation
  • AI document structuring via OCR and LLM
  • Parallelization of OCR and structuring pipelines
  • CI/CD deployment on Azure DevOps with Docker and monitoring

Skills

Publications

• Federated Learning for UAV Perception Task: A Survey
  TechRxiv (under review)

  Survey on federated learning approaches for UAV (drone) perception tasks. Preprint under review.

  View Publication

Teaching

• Artificial Intelligence

  2025
  École Centrale d'Electronique, Paris

  Role: Professor - Engineering cycle (bac+5)
• MLOps

  2025
  École Centrale d'Electronique, Paris

  Role: Professor - MSc AI (bac+5)
• Mathematics

  2025
  École Centrale d'Electronique, Paris

  Role: Professor - Preparatory class (bac+1)

Portfolio

• NASDAQ Closing Price Prediction
  Portfolio

  Time series model for predicting NASDAQ closing price. Kaggle silver medal, top 4%. [PyTorch, LightGBM, k-Fold]

  View Project
• ASL to Text Video Translation
  Portfolio

  Model to transcribe American Sign Language (ASL) from video to text. Kaggle competition by Google, top 4%. [TensorFlow, TFLite, MediaPipe, Conformers, OpenCV]

  View Project
• Real-time CV Search Platform
  Portfolio

  Production platform for real-time CV search based on semantic retrieval. [Redis, Docker, Weaviate, ETL, Grafana, Kubernetes, Amazon S3]

  View Project

Languages

• French

  Native
• English

  B2
• Spanish

  B1
• Arabic

  A2