I am a third-year engineering student at ENSTA Paris and a Master’s student in Data Science at École Polytechnique (IP Paris), pursuing a dual degree through the academic partnership with:

• ENIT – National Engineering School of Tunis
• ENSTA Paris – École Nationale Supérieure des Techniques Avancées

My background combines applied mathematics, probability, statistics, and optimization with advanced data science and machine learning, with a strong focus on real-world, high-impact applications.

• 📊 Statistical Learning & Data Science
• 🤖 Machine Learning & Deep Learning (MLP, CNN, RNN, Transformers)
• 🧠 NLP & Large Language Models (RAG, fine-tuning, prompt engineering)
• 📈 Time Series & Spatio-Temporal Modeling
• 🔗 Extreme Value Theory & Dependence Modeling
  (Copulas, Pickands dependence function, spatial extremes)
• 🌍 Climate, Risk & Energy Data

I am particularly interested in projects that connect theoretical modeling with large-scale, real-world data, emphasizing interpretability, robustness, and high-quality visualizations.

• Languages: Python, R, SQL, C/C++, Bash
• ML / DL: Scikit-learn, PyTorch, TensorFlow, Optuna
• NLP & LLMs: Transformers, LangChain, RAG, LoRA, Quantization, SFT / DPO
• Data & Visualization: Pandas, NumPy, Matplotlib, Seaborn, Power BI
• Scientific Computing: SciPy, Statsmodels, RStudio
• Tools: Git, GitHub, Docker (basics), Linux, LaTeX

Neural estimation of the Pickands dependence function for Gumbel, Galambos, and Tawn copulas, including convex projection, simulation studies, and applications to spatial climate data.

Spatio-temporal modeling of meteorological and climate data (Météo-France), focusing on extreme events, dependence structures, and risk measures (VaR, TVaR).

Design of a Retrieval-Augmented Generation (RAG) system with dynamic knowledge bases, traceable answers, and interactive dashboards using LangChain, ChromaDB, and Streamlit.

End-to-end ML pipeline handling highly imbalanced data, feature selection, model comparison (XGBoost, Random Forest), and robust evaluation (ROC-AUC, F1-score).

Exploratory analysis and predictive modeling of fuel consumption (mpg) using the mtcars dataset: correlation analysis and visualization, multiple linear regression, stepwise AIC variable selection, and PCA + Principal Component Regression (PCR), with model comparison using ANOVA.

• Develop robust, interpretable, and scalable ML models
• Work on research-driven or high-impact applied data science projects
• Contribute to domains such as climate, risk, energy, and AI systems
• Maintain high standards in code quality, reproducibility, and visualization

• GitHub: you are here 😊
• LinkedIn: Rayen Zargui
• Email: rayen.zargui@ensta-paris.fr