MSc Robotics graduate at TU Delft. Passionate about path planning, computer vision, and visual language action models.
>> Loading modules...
>> TensorFlow initialized.
>> Kinova_Driver ready.
_
I am an MSc Robotics graduate from the Technical University of Delft, where I specialized in path planning, computer vision, and visual language action models. I also hold a BSc in Mechanical Engineering from the Technical University of Eindhoven.
I am also highly motivated in the Deep Learning and AI fields, constantly seeking to expand my knowledge and apply it to real-world robotics challenges. I am always trying to apply anything I learn in PyTorch to understand the architectures.
Graduated with a 7.3 GPA and an 8.5 for the Bachelor Thesis. Extra Course/Exchange: Cybernetics & Artificial Intelligence at Technical University of Prague.
Achieved a 7.5 GPA. Relevant coursework: Machine Learning (TensorFlow, Keras, Scikit-Learn, PyTorch), Path Planning (MPC, RRT*).
Specializing in Path Planning using RRT* for global path and MPC for local obstacle avoidance. Placed 5th in the 2024 Eindhoven B division.
Benchmarked several 3D SLAM and localization packages. Implemented full coverage path planning for a metro cleaning robot using Nav2.
Explored VLA models Octo and NVIDIA Gr00t. Integrated LLM to introduce reasoning. Used Franka Emika FR3.
Developing learning based methods for manipulators using force sensors. Implementing various robot control algorithms, such as admittance and impedance control.
Implementation of RRT* for global planning and Model Predictive Control (MPC) for local obstacle avoidance in dynamic environments.
Global planner implementation using Dijkstra algorithm for a quadrotor in PyBullet simulation, ensuring safe maneuverability.
Deep Learning reproduction study classifying malware images using CNNs with nested cross-validation and hyperparameter tuning.
Development of a robust autonomous navigation system using Nav2, Fast-LIO SLAM, and coverage planning for metro cleaning.
Integrating Large Language Models with Visual-Language-Action models (GR00T) to improve reasoning for long-horizon robotic tasks.
A complete implementation of the Vision Transformer (ViT) architecture in pure PyTorch, trained on CIFAR-10.