I'm passionate about AI and I'm currently learning about it. I like implementing Deep Learning models and I'm also interested in Reinforcement Learning.

Check out my Medium profile for more details: Maxime Szymanski on Medium

• Single and Multi-Agent Reinforcement Learning Approach to - Optimize Aircraft Ground Trajectories at Airports: This paper is about a reinforcement learning approach to optimize aircraft ground trajectories at airports. It is implemented in Python and is tested on a simplified airport environment.

  Relevant features : PPO, Single and Multi-Agent Reinforcement Learning.

• Development of a Map-Matching Algorithm for the Analysis of Aircraft Ground Trajectories using ADS-B Data: This paper is about a map-matching algorithm for the analysis of aircraft ground trajectories using ADS-B data. It is implemented in Python and is tested on real aircraft trajectories.

  Relevant features : Hidden Markov Model, Viterbi algorithm.

• DeepRL Atari implementation : This project is a collection of Deep Reinforcement Learning algorithms. It is implemented in PyTorch and is tested on OpenAI Gym environments, including AtariGames.

  Relevant features : DQN, PPO, DDPG, AtariGames, Pytorch.

• PPO implementation : This project is a self-made PPO, a reinforcement learning algorithm. It is implemented in PyTorch and is tested on OpenAI Gym environments.

  Relevant features : Discrete and Continuous action spaces handling, TensorBoard logging, Hyperparameters tuning, etc.

• SplendIA : AI to play Splendor board game : This project is an AI to play Splendor board game. It is implemented in Python and is tested on Splendor board game. As a team, we implemented the board game and the AI. I was in charge of the AI part. We tested Monte-Carlo Tree Search and Alpha-Beta pruning algorithms. Moreover, we implemented a PPO algorithm to play the game, as a Self-Play algorithm.

  Relevant features : Self-Play reinforcement learning, Alpha-Beta pruning, Monte-Carlo Tree Search, etc.

• Stocks Clustering : This project is a stocks clustering algorithm. It is implemented in Python and is tested on historical stocks data.

  Relevant features : K-Means clustering, PCA, Auto-Encoder, etc.

• Languages: Python, C, C++, Java
• Deep Learning: PyTorch, TensorFlow, Keras
• Machine Learning: Scikit-learn, Pandas
• Other: Git, Linux, Docker, SQL