
Portfolio
Christian Remmelberger
Computer Science Student at Technical University Munich. Focusing on the intersection of Robotics, AI, and Human-Machine Interaction. Currently exploring teleoperation and reinforcement learning for dexterous robotics. Formerly at Microsoft and Agile Robots.
About
Education & Research Focus
Computer Science student at TUM, supported by the best.in.tum excellence scholarship. Exchange semester at UC San Diego focused on Computer Vision and NLP. Research interests in robotic manipulation, AI, and humanoids.
Professional Experience
Student Researcher at the TUM Chair of Learning AI, building teleoperation systems for dexterous robot hands. Former intern at Microsoft (deepfake detection & signal analysis) and Agile Robots (teleoperation software for humanoids).
Projects & Initiatives
Leading EyeVoices, an eye-blink-based communication platform for people with severe paralysis. Robotics engineer at RoboTUM, training bipedal robots via reinforcement learning. Working primarily with Python, C++, and PyTorch.
Education
Technical University of Munich (TUM)
B.Sc. Computer Science
Munich, Germany
Oct 2022 — Present
Recipient of the best.in.tum excellence scholarship, awarded to the top 2% of the Informatics Department. Focus on machine intelligence and mathematics. Current GPA: 1.3 (best 1.0, pass 4.0, fail 5.0).
University of California, San Diego (UCSD)
Exchange Semester
San Diego, CA, USA
Sep 2024 — Dec 2024
UPS Extended Studies Program with specialized coursework in Natural Language Processing, Computer Vision, and Robotics.
Work Experience
TUM Chair of Learning AI for Dexterous Robots
Student Researcher
Munich, Germany
Nov 2025 — Apr 2026
Developing a teleoperation system for a dexterous robot hand. Implementing a learning-based retargeting model that maps human hand motions to the robot's distinct kinematics for precise in-hand manipulation.
Microsoft
Software Engineering Intern
Dublin, Ireland
Aug 2025 — Oct 2025
Developed honeypot systems to detect and analyze malicious signals within MSTIC. Fine-tuned EfficientNet-B0 (timm) for deepfake detection with a custom GAP/Dropout/Linear head and Penrose pooling layer for enhanced feature aggregation. Honorable Mention at Microsoft Global Hackathon for 'Blink+'.
Agile Robots
Robotics Trainee
Munich, Germany
Mar 2024 — Jul 2024
Developed teleoperation software for humanoid robotic systems using Python, C++, and ROS, enabling high-quality demonstration data for downstream VLA and imitation-learning pipelines. Integrated GELLO, haptic gloves, and trackers for precise and responsive manipulation. Delivered live demonstrations to partners and customers.
Selected Projects


Contact
I'm always open to discussing new research collaborations, engineering challenges, or interesting opportunities. Reach out at christian.remmelberger@tum.de
