About Stephane
"Uncovering order in complexity — ✨ Innovative solutions for data-driven success 📊"
What Sets Me Apart 🏆
My Services 🌟
- Data Visualization: Designing clear and aesthetic dashboards to make complexity accessible.
- Modeling and Simulation: Creating advanced models to optimize your decisions.
- Data Analysis and Transformation: Strategically leveraging data through modern techniques like machine learning.
Technical Skills 🔧
- Data Analysis and Extraction: Big Data, SQL/NoSQL, Data Mining.
- Advanced Modeling and Simulation: Dynamic and stochastic methods.
- Supervised and Reinforcement Learning: TensorFlow, PyTorch, OpenAI Gym.
- Data Visualization: Dash, Plotly.
- Technical Environment: Python, C++, Linux, Git, HPC clusters.
French
Native or bilingual
English
Native or bilingual
Experience
- DotworldData scientistE-COMMERCEApril 2024 - Today (2 years and 4 months)a) Customer classification based on chargeback risk for a mid-provider
- Data preprocessing and customer geolocalization to analyze regional behaviors.
- Encoding categorical features (devices, browsers, currencies) using one-hot and frequency encoding.
- Development of a classification algorithm based on SVM and XGBoost to estimate chargeback probability.
Result: Significant reduction in chargeback costs through precise customer allocation based on risk score.b) Reinforcement learning-based A/B testing- Definition of the reward function to optimize feature tests.
- Use of a "classic" reinforcement learning algorithm based on UCB, accelerated by a neural network to approximate the reward function.
Result: Improvement of A/B tests and optimization of tested features. - King's College LondonPostdoctoral Researcher -- Marie Curie FellowRESEARCHApril 2021 - Today (5 years and 4 months)Kashiwa, Japan- Simulation of particle detectors and design of classical and Machine Learning algorithms (Python, C++, Fortran). Contributor to an open-source physics software (WCSim).- Management of research grants (280k EUR, 450k JPY) to design a photosensor test bench (mechanical, electronic, readout). Organization of the review of a 20+ M USD project for the installation of photosensors in Hyper-Kamiokande.- Setup of a neutrino physics laboratory, including design, procurement, and documentation. Automation of data collection and analysis (Python, C++, ROOT) and deployment of programs on computing grids.- Comparison of the performance of photosensors from 4 different manufacturers, measurement of their characteristics for certification, and ongoing negotiation with manufacturers.
- Genius sportsConsultant -- Quantitative AnalystENTERTAINMENT AND LEISUREOctober 2021 - January 2023 (1 year and 3 months)- Prediction of arrival time for horse races using a gamma distribution model in Python and C++- Time series forecasting for horse race rankings, using TensorFlow, scikit-learn, and pandas- Construction of a reinforcement learning environment for betting and its challenges, using Keras and OpenAI Gym
Reviews
Recommendations
Be the first to recommend Stephane
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Engineering DegreeGrenoble-INP Phelma2013Diplôme d’Ingénieur en Physique et Nanosciences
- Master of ScienceImperial College London2013MSci de l'Imperial College London en Quantum Fields and Fundamental Forces