About Me
I am pursuing my PhD in artificial intelligence with a focus on reinforcement learning. With my research, I aim to enhance the capabilities of neural agents in complex environments. As a member of the Research Training Group “Neuroexplicit Models of Language, Vision, and Action”, I am particularly interested in models that combine neural and explicit approaches to leverage the strengths of both. My current research is on improving exploration strategies in reinforcement learning. Effective exploration strategies can significantly increase the sample efficiency, robustness, and safety of neural agents, enabling them to learn previously intractable tasks.
Interests
Besides work, I enjoy tinkering with technology, such as customizing my system and homelab, improving my workflow and dotfiles, working on personal projects like my website. I also take photos, which I (very irregularly) post there.
To balance things out, I do a variety of sports: I go to the gym, run, and boulder. When I started my PhD, I developed an interest in learning languages. I started with Spanish and Japanese, but now I am only learning Spanish, currently at B2 level.
- Artificial Intelligence
- Reinforcement Learning
- Computer Graphics
- Advanced Sampling Methods
Education
PhD in Computer Science (Artificial Intelligence), ongoing
Co-supervised by Prof. Verena Wolf and Prof. Philipp Slusallek at Saarland University
MSc in Computer Science (Computer Graphics), 2024
Supervised by Prof. Philipp Slusallek at Saarland University
MARS: Multi-sample Allocation through Russian roulette and Splitting
Grade: 1.0 (overall 1.0)
BSc in Computer Science (Computer Graphics), 2022
Supervised by Prof. Philipp Slusallek at Saarland University
Integrating a Denoising Neural Network Into a Renderer Using a High-Level Functional Language and Partial Evaluation
Grade: 1.0 (overall 1.5)
Publications (3)
2026
PGTG: Procedurally Generated Grid-Based Traffic Gym
2026 — Reinforcement Learning Journal
2025
PyDSMC: Statistical Model Checking for Neural Agents Using the Gymnasium Interface
2025 — Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems - Second International Joint Conference, Proceedings. International Joint Conference on Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems (QEST+FORMATS-2025), August 26-28, Aarhus, Denmark
2024
MARS: Multi-sample Allocation through Russian roulette and Splitting
2024 — SIGGRAPH Asia 2024 Conference Papers. ACM SiggraphAsia (SigAsia-2024), December 3-6, Tokyo, Japan
Teaching (4)
Special Topics in Deep Reinforcement Learning
Assistant (English), in Summer 2026
Seminar at Saarland UniversityArtificial Intelligence
Tutor (English), in Winter 2023/24
Saarland UniversityRealistic Image Synthesis
Tutor (English), in Summer 2023
Saarland UniversityHöhere Mathematik für Ingenieure 1
Tutor (German), in Winter 2020/21
Saarland University
Supervised Students (10)
Behkam Fallah
Master's Thesis, ongoing
Nils Waltner
Research Assistant, ongoing
Felix Maurice Kuntz
Research Assistant, ongoing
Behkam Fallah
Research Assistant, ongoing
Nils Waltner
Learning Safer Multimodal Diffusion Policies Using Latent Representations
Bachelor's Thesis, 2026
Ruslan Dontsov
Learning Inherently Interpretable Reinforcement Learning Policies
Bachelor's Thesis, 2026
Moritz Oster
Research Assistant, 2026
Philipp Sauer
Classification and Analysis of Safe Reinforcement Learning Algorithms
Master's Thesis, 2025
Michel Scherer
Lifting the RARE Algorithm to Continuous State and Action Spaces
Bachelor's Thesis, 2025
Ivo Hoese
Research Assistant, 2025