Welcome to my GitHub profile! I'm a passionate Data Science MSc student at École Polytechnique and an engineering student at ENSTA Paris. With a strong academic background in Machine Learning, Deep Learning, and Applied Mathematics, I enjoy tackling real-world challenges through innovative data-driven solutions.
Throughout my journey, I have applied my skills across diverse fields, embedding and augmenting heterogeneous graphs to predict insurance premium amounts. This has allowed me to leverage complex relationships in data for impactful predictions and solutions.
I thrive on collaboration, adapting quickly to new environments, and bringing creative ideas to life. I’m always eager to explore cutting-edge technologies and contribute to impactful projects that make a difference.
Currently, I am seeking a 6-month final-year internship starting in April 2025 to apply my expertise in Data Science to innovative projects. Additionally, I am an active Kaggle competitor, working hard to become an expert by tackling challenging data science problems and contributing to the community, being ranked among the top 2% in consecutive challenges.
- 🎓 Educational Background: Currently pursuing a master's degree in Data Science at École Polytechnique and finishing my third year in engineering at ENSTA Paris with a specialization in Science of Data and Optimization.
- 🌟 Current Focus: Working on projects related to different fields, from Large Scale ML problems, Time Series, and Deep Learning to Graph Neural Networks, NLP, and LLMs.
- 💡 Interests: Exploring advanced topics in Data Science, Machine Learning, AI ethics, and building innovative side projects that combine creativity and technology.
Role: Intern | Organization: Telecom SudParis
- Implemented various algorithms, including Relation Graph Convolution Network (R-GCN), Heterogeneous Graph Transformer (HGT), and Graph Attention Transformer (GAT).
- Performing heterogeneous graph data augmentation using graphon.
A collaborative effort to develop a predictive model for groundwater levels, addressing environmental concerns and sustainable water resource management.
- Filled the missing values using a sliding window approach.
- Designed a predictive model for groundwater levels in multiple departments in France using the AutoGluon library with a focus on LightGBM and FastAI models.
- Fine-tuned the hyperparameters using the Optuna library.
Rank: 51/2390
- A project aimed at predicting insurance premium amounts by leveraging advanced machine learning models and embedding techniques.
Participation in a Kaggle competition focused on predicting credit default risks for financial portfolios.
- Performed a data Analysis and statistical tests to determine which variables are more explicative.
- Train multiple classification models: XGBOOST, LighGBM, CatBoost etc..
- Participation in a Kaggle competition focusing on innovative approaches to graph generation.
- Deploy OpenCv library in python to detect objects.
- Built an interactive app using C++ and Python to calculate optimal trajectories in obstacle-filled environments with Dijkstra’s Algorithm.
- 🏆 Hackathon Enthusiast: Regular participant in hackathons to push boundaries and test creative solutions.
- 🎖️ Kaggle Competitor: Ranked among the top 3% in two consecutive competitions.
- 📚 Learning: Constantly improving my knowledge in AWS, PowerBI, and advanced visualization tools.
- 💬 Ask Me About: Data Science, Graph Neural Networks, and AI applications.
Thanks for stopping by! Feel free to explore my repositories and get in touch if you'd like to collaborate. 😊