Flim de Jong
Robotics Student (Software & AI) //
University of TwenteAbout
Robotics Engineer (MSc) | Advancing Embodied AI through the use of Foundation Models (VLM/VLA) and RL to enable real-world reasoning and autonomy.
Projects
Robotic Foundation Models (VLA)
Fine-tuned the SmolVLA (0.45B) model for physical 6-DoF manipulation. Collected custom teleoperation datasets and implemented a BERT-based intent classifier to bridge language and action.
Next Best View Planner
Developed a DQN-based planner for autonomous 3D reconstruction. Outperformed greedy approaches by 10-30% using a hybrid RANSAC/ICP pipeline and RealSense RGBD data.
Autonomous Robot Soccer (RoboTeamTwente)
Guided the technical roadmap for an autonomous robot soccer system across a multi-disciplinary team of 20 members. Implemented high-level strategy using FSMs and low-level control using Reinforcement Learning (PPO) in Mujoco MJX.
Experience
Technical Manager & Software Engineer | RoboTeamTwente
Managing a multi-disciplinary team of 20 to develop autonomous soccer robots. Focused on real-time AI strategy and RL-based motion control.
Education
MSc in Robotics (Software & AI)
University of Twente, NL. Specializing in Algorithms and Software AI with a focus on Deep Learning, Foundation Models, and autonomous robot perception.
BSc in Advanced Technology
University of Twente, NL. Multidisciplinary foundation in Mechatronics, applied physics and dynamical systems, with electives in ML during an international exchange.