CV
Education
Ph.D in Machine Learning
Autonomous Learning Robots @ Karlsruhe Instite of Technology (KIT), 2020 - 2025 (expected)
- World modeling and representation learning for reinforcement learning from multimodal observations
- Uncertainty in deep state space models for model‑based reinforcement learning and Bayesian methods
- Combining recurrent neural networks and Kalman filters for uncertainty modeling in sequential data
- Variational inference for time‑series, meta‑learning, and reinforcement learning
Master of Science in Autonomous Systems (Computer Science)
Technische Universität Darmstadt, 2016-2019
- with distinction
- Thesis: Expected Information Maximization: Using the I‑Projection for Mixture Density Estimation
Bachelor of Science in Computer Science
Technische Universität Darmstadt, 2014-2016
- Thesis: Learning Deep Feature Spaces for Nonparametric Inference
Summer Schools
- Nordic Probabilistic AI School, Trondheim, Norway, June 2023
- Machine Learning Summer School, Virtual (Tübingen, Germany), July 2020
Work experience
Doctoral Researcher
Karlsruhe Institute of Technology (KIT), Autonomous Learning Robots, since 2020
- Supervision of 25+ student theses and projects (Full List).
- Head teaching assistant over 6 semesters for the lecture “Machine Learning - Foundations and Algorithms”
- Guest lectures on model-based reinforcement learning at KIT
- Setup and administration of IT infrastructure, including Slurm-based HPC
Research Scientist
FZI Research Center for Information Technology, (FZI Forschungszentrum Informatik), since 2023
- Establishing a new research group under the direction of Prof. Gerhard Neumann
- Grant writing for European and German national research calls with academic and industrial partners.
- Structured reinforcement learning for control applications of an industrial customer
Doctoral Researcher
Bosch Center for Artificial Intelligence, Robert-Bosch-GmbH, 2019
- Research on versatile imitation learning
- Followed advisor to KIT after 8 months
Student Research Assistant
Technische Universität Darmstadt, Intelligent Autonomous Systems, 2016-2019
- Various manipulation robotics projects on real robots and in simulation
Internship
Volkswagen Group, Data Lab, 2017
- Deep Learning and Robotics Challenge: Sorting Lego bricks using a Lego robot controlled by deep learning
Reviewing
Machine Learning
- Conference on Neural Information Processing Systems (NeurIPS)
- International Conference on Machine Learning (ICML)
- Conference on Robot Learning (CoRL)
- Reinforcement Learning Conference (RLC)
Robotics
- IEEE Robotics and Automation Letters (RA-L)
- International Conference on Intelligent Robots and Systems (IROS)
- International Conference on Robotics and Automation (ICRA)
Skills
- Languages: German (native), English (fluent)
- Programming: Python, Java, C++, C, Matlab
- ML Libraries: Pytorch, Tensorflow, Numpy, Scipy, Optuna, Hydra, Scikit-Learn, Jax
- Tools: Linux, Git, Slurm, LaTeX, Microsoft Office
- Robotics: Mitsubishi PA10, KUKA LBR iiwa, Franka Emika Panda
Student Supervision
I supervised > 20 Bachelor and Master Students during my time at the Karlsruhe Institute of Technology
- Ralf Herp - Bachelor Thesis - Diffusion-based Imitation Learning for Contact Rich Manipulation
- Erwin Müller - Master Thesis - Model-based Reinforcement Learning of Diverse Skills
- Keven Sahin - Master Thesis in cooperation with Robert Bosch GmbH - Offline Reinforcement Learning for Hydraulic Excavator Control
- Sven Hillenkötter - Mester Thesis - Model-Based Reinforcement Learning for Challenging Manipulation Tasks
- Yasin Tatar - Master Thesis - Investigating the Amortization Gap in Variational Inference for State Space Models
- Claudius Kienle - Master Thesis in cooperation with ArtiMinds Robotics GmbH - Multimodal Transformer for Zero-Shot Optimization of Robot Manipulation Tasks Across Environments
- Ruben Jacob - Master Thesis - Episodic Reinforcement Learning with Transformer Models
- Aina Galofre - Master Thesis - Improving Uncertainty Estimation for Model-Based Reinforcement Learning
- Christian Schorr - Master Thesis -Improving the Soft Actor-Critic with Information Theoretic Trust Regions
- Florian Seligmann - Bachelor Thesis \& Research Project - Benchmarking Modern Bayesian Deep Learning Algorithms
- Stefan Krieg - Bachelor Thesis - Tackling Offline Reinforcement Learning by Model Learning
- Sebastian Markgraf - Master Thesis - Self-Supervised Model-Based Reinforcement Learning under Partial Observability
- Maximilian Li - Master Thesis - Imitation Learning for Versatile Robot Manipulation
- Ahmed Agha - Master Thesis - Improved Trust Regions for Adversarial Imitation Learning
- Moritz Reuss - Master Thesis in cooperation with Robert Bosch GmbH - Hybrid Inverse Dynamics Models for Robot Impedance Control
- Giao Nguyen Quynh - Master Thesis in cooperation with Robert Bosch GmbH - Kalman Mixture Networks
- Philipp Dahlinger - Research Project \& Master Thesis - Information Theoretic Trust Regions for Gradient Descent
- Roman Freiberg - Bachelor Thesis - Actor-Critic Methods for Robotics
- Niklas Freymuth - Master Thesis - Inverse Reinforcement Learning for Highly-Versatile Behavior
- Carmen Eitel - Master Thesis - Deep Reinforcement Learning under Partial Observability using Kalman Filtering