I'm Jelle Jansen, a skilled Data Scientist located in Heerlen, The Netherlands. My expertise in Deep Learning and ML/AI Ops shapes my advanced analytical skills and strategic problem-solving capabilities. What fuels my passion is the opportunity to build something with a team that is impactful and make it available to everyone, providing valuable support in their daily lives.
My fascination with AI, sparked by its gaming capabilities, led me to pursue a Master's in Artificial Intelligence. This passion, combined with my quick learning skills and enthusiasm for acquiring new knowledge, motivates me to keep up with the latest developments in the field of AI.
View ResumeAs part of my master thesis, I designed a search algorithm that represents a variation of Monte Carlo Tree Search. The algorithm utilizes implicit minimax backups in conjunction with a neural network to assess the values of various game states. By training the neural networks with the descent framework, the algorithm was able to outperform other state-of-the-art search algorithms in the game of Breakthrough.
As part of a group assignment for my Master's in AI, I designed a layer for a Neural Network that represents Mental Models. Since the results were promising, our work has been published as a conference paper for the Benelux Artificial Intelligence Conference.
My very first project related to AI, was an implementation of the well-known Chrome T-Rex Dinosaur Game in pygame, employing learning through the NEAT-python package. Despite imperfections in the game, code, and dinosaur performance, this experience showed me what the potential of AI. It motivated me to pursue a career in this field.
Drawing inspiration from the Deep Reinforcement Learning techniques explored in my MSc. Thesis and the game development experience from my (First) Dino Game, I embarked on a project to create a racecar capable of navigating circuits autonomously using Neural Networks and Proximal Policy Optimization for training.
As part of an individual assignment for my Master's in AI, I implemented Alpha-Beta search algorithms with numerous enhancements for the game Cannon. This experience not only allowed me to compete effectively with fellow students in a tournament, but also served as a significant source of inspiration for pursuing my master's thesis in this field of research.
In anticipation of an upcoming assignment, I developed a solid foundation by implementing diverse search algorithms employing self-play to improve their efficacy in playing Checkers. This code proved to be a valuable groundwork for the subsequent assignment.
As part of a group assignment for my Master's in AI, I developed a Convolutional Neural Network architecture to achieve an optimal performance on the Facial Expression Recognition dataset, by iteratively optimizing the parameters and training process.
As part of a group assignment for my Master's in AI, I implemented a simulation that illustrates the uncertainties faced by robots in real-world scenarios with respect to their actual position, arising from factors such as sensor quality (measurement noise) and uncertainty in the motion model.
As part of a group assignment for my Master's in AI, I developed a circular vacuuming robot designed to efficiently gather dust. This robot employs a genetic algorithm, utilizing its 12 sensors along with a Recurrent Neural Network (RNN) to calculate and adjust the velocities of its left and right wheels, aiming to optimize the dust collection process.
As part of a group project for my Premaster's in AI, I implemented multiple search algorithms for the game Backgammon. The search algorithms varied random bots to algorithms trained with the Genetic Algorithm and self-learning.
Let's connect and create something amazing together!