I am a passionate AI researcher and entrepreneur specializing in Model-Based Reinforcement Learning and World Models. Currently pursuing my MS in AI (cum laude) at the University of Amsterdam as an ELLIS Honours student, while co-founding Traize and conducting RL research at TU Munich.
Key Highlights
🎓 MS in AI (cum laude) - University of Amsterdam
🏆 UvA ELLIS Honours Program Student
🚀 Co-Founder & Tech Lead at Traize
🔬 Research Assistant RL - TU Munich
💼 Lead AI Engineer - Infinitas Security
📚 Triple Bachelor's: CS, Economics & Mathematics
My Journey
2020 Bachelor of Science Economics
2022 KPMG Machine Learning Engineer
2022 Bachelor of Science Information Systems
2023 Bachelor of Science Mathematics
2024 Research Assistant NLP University of Passau
2024 Master of Science Artificial Intelligence (Honours)
2024 Lead AI Engineer Infinitas Security
2025 Research Assistant RL University of Bielefeld
2025 Co-Founder & Tech Lead Traize
2025 Research Assistant RL TU Munich
Education
I am a Master's student in Artificial Intelligence (cum laude) at the University of Amsterdam, where I am also an Honours Student in the UvA ELLIS Honours Program. I hold dual Bachelor's degrees in Information Systems, Economics from University of Passau. I am also pursuing a Bachelor's in Mathematics at Fernuniversität Hagen at the moment. My education provides a strong foundation in research, engineering, and mathematics for AI research.
Research
My core research focus is on Model-Based Reinforcement Learning and World Models. I am particularly interested in making learning in and with world models more efficient through curriculum learning, exploration strategies, and uncertainty estimation for policy learning. I am currently collaborating with Max Planck Institute Tübingen and UvA on developing a curriculum learning-based world model agent for model-based RL.
Beyond world models, I work on Self-Supervised Learning for vision and language, applying this knowledge to create language-grounded world models. I also use RL to solve combinatorial optimization problems, leveraging self-supervised learning to create problem representations that enable more efficient learning with model-based RL approaches. I am currently collaborating with labs at University of Vienna and TU Munich on these topics.
As a Research Assistant at TU Munich, I develop novel representation learning and reinforcement learning approaches for combinatorial optimization under uncertainty. My work combines graph neural networks, self-supervised learning, and model-based RL to tackle complex decision-making problems.
Recent work: Developed DecoDINO for 3D human-scene contact prediction (CVPR 2025 RHOBIN Challenge, +7% F1 improvement), contributed to a GNN interpretability reproduction study (accepted to TMLR), extended DreamerV3 with intrinsic rewards for sparse-reward exploration, and developed GNN-based RL approaches for data center optimization. I actively implement state-of-the-art research papers and share insights through blogs and open-source contributions.
Professional Experience
I am currently Lead AI Engineer at Infinitas Security (since March 2024), leading all AI features including LLM inference pipelines and custom graph-based retrieval systems using state-of-the-art MLOps methods. Since May 2025, I am also Co-Founder & Tech Lead at Traize (traize.ai), an intelligent operating system for enterprise knowledge, where I design and implement the end-to-end microservice-oriented backend and train proprietary explainable GNN-based retrieval algorithms using FastAPI, Kubernetes, and agent-oriented workflows.
Previously, I worked as a Machine Learning Engineer at KPMG (April 2022 - 2025), where I built customized center-point object detection models for satellite images, developed full-stack applications with React and FastAPI, and contributed to internal LLM-based RAG systems.
On the research side, I am currently a Research Assistant at TU Munich (since 2025), developing novel representation learning and reinforcement learning approaches for combinatorial optimization. Previously, I was a Research Assistant at University of Bielefeld (2025 - November 2025) working on reinforcement learning, and at University of Passau (March 2024 - May 2025), where I focused on self-supervised pretraining of domain-specific language encoder models and developed modular pretraining parameter tuning frameworks.
A list of papers I have read and find interesting. Mainly RL and Computer Vision based papers.
I've worked on diverse projects spanning AI, Machine Learning, and Software Development, focusing primarily on Natural Language Processing, Computer Vision, and Reinforcement Learning. Additionally, I'm engaged in Full Stack Development projects. Feel free to check them out :)