Dong Tian

Hamburg University of Technology (TUHH)
Harburger Schloßstraße 28, Channel 4 (Building HS28)
21079 Hamburg
Room: 2.016

Artificial Intelligence for Enhancing Operation and Exploitation of X-ray Free-Electron Lasers (AIOPs4XFEL)
Supervisors: Prof. Kai Rossnagel (DESY), Dr. Markus Scholz (DESY), Prof. Olaf Landsiedel (TUHH)

Dong is a doctoral researcher with an M.Sc. in Mechatronics and Information Technology from Karlsruhe Institute of Technology (KIT), where he specialized in machine learning, optimization, probability, control theory, and algorithmic modeling. His Master’s thesis, “Multi-Step Q-Learning with Transformer-Based Critic,” developed transformer-based critic models for multi-step, off-policy reinforcement learning and contributed to peer-reviewed work at ICLR. His research experience includes transformer-based reinforcement learning, off-policy optimization, simulation-based learning, and sequence modeling, including a first-author ICLR 2026 paper and a second-author ICLR 2025 Spotlight paper. He also contributes to the machine-learning research community as a main track reviewer for NeurIPS 2026.
From March 2023 to September 2024, he worked at Porsche Engineering Group GmbH as an intern, focusing on robotics-related development, hardware-in-the-loop testing, simulation and automation, and computer vision. Since June 2026, he has been working at DESY on machine learning and edge AI methods for near-real-time XFEL beam diagnostics, feedback, and optimization. His work focuses on AI-based control for beam tuning, data quality, and experimental efficiency under accelerator-physics constraints. He is also interested in the mathematical and computational foundations of physical systems.