French, 22 years old

Raphaël Boige

Research intern in Reinforcement Learing at Laboratoire Cedric (CNAM).


2020-2021 Ecole Polytechnique, Paris (Palaiseau) Master Data Science (applied mathematics department).

2018-2021 Télécom Paris, Paris (Palaiseau) Graduate engineering school, majoring in computer science and applied mathematics. (GPA 3.93/4.00)

2016 - 2018 Lycée Jean-Baptiste Say, Paris Classes préparatoires

2013-2016 Lycée Albert Camus, Fréjus Baccalaureate with highest honours


2021 Research Intern in the computer science department of Conservatoire National des Arts et Metiers (CNAM). We try to generate and quantify diversity among reinforcement learning agents. We focus on the notion of perceptual diversity and we link it to the experience of a video game player. Keywords : Deep Reinforcement Learning, Intrinsic Motivation, Quality-Diversity, Skill discovery.

2020 Co-author of the book Intégrer l’X ou l’école d’ingénieurs de tes rêves avec les Sherpas, published in June 2020 by Vuibert (Albin-Michel). The book is a guide for undergraduate students to suceed at the nationwide competitive exams for engineering schools.

Academic Projects

2021 Developer of a project on pretraining parameters of a Deep Neural Network using a Deep Belief Network. Implementation from scratch in Python of the DBN structure and CD-1 algorithm. Comparison of the performance between the randomly initialized and the pretrained network on the MNIST dataset.

2020 Winner of the internal Data Challenge of the Msc on Magnetoencephalography (MEG) source localization. The solution used a custom implementation of adaptive Lasso for sparse support recovery and a meta-algorithm for fine-tuning the parameters.

Personal Projects

2021 Finalist of the 2021 Huawei Challenge on transfer learning for detecting failures of optical networks. Time series analysis and domain adaptation using both standard statistical tools and deep learning methods.

2021 Competitor in the 2021 edition of Hackaton Hi!Paris (Organized by Ecole Polytechnique and HEC Paris). Optimization of the energy consumption of a building with solar panels using a smart grid controller. We leveraged reinforcement learning methods as well as heuristics to minimize the cost and the environemental impact of our building.

2020 Maintainer of an experiment on predicting outcomes of competitive Counter-Strike games, including a theoretical approach (probabilistic and gambling theory), an implementation using different elo-based computed features as well as simulations to validate winning strategies.


2019-2020 Vice-President of the students sports office an association with 400 contributors, 13 members and more than 80 k€ of budget.

2019-2020 President of the cheese club Téléfrom

Skills & Interests

Programming languages: Python (pandas, sklearn, pytorch, gym). A bit of Java, C++, C, JS. Tools: Git, Jupyter. Languages : French native speaker, working profiency in English (C1).