Scholar
Jan-Nico Zaech
Google Scholar ID: suiZThYAAAAJ
Research Scientist, INSAIT, Sofia University
Citations & Impact
Citations
846
H-index
13
i10-index
15
Publications
20
Co-authors
17
list available
Publications
Resume (English only)
Academic Achievements
- Publications: 'GaussianVLM: Scene-centric 3D Vision-Language Models using Language-aligned Gaussian Splats for Embodied Reasoning and Beyond' (arXiv, July 2025); 'ReVLA: Reverting Visual Domain Limitation of Robotic Foundation Models' (ICRA 2025, February 2025); 'Occam’s LGS: An Efficient Approach for Language Gaussian Splatting' (arXiv, December 2024); Patent: X-ray machine learning model based on master's thesis work granted; Conference Presentations: Articulate3D accepted at ICCV 2025; Presented ReVLA at ICRA 2025.
Research Experience
- Currently a research scientist at INSAIT, leading a team on various aspects of foundation models for robotic applications covering end-to-end control and embodied scene understanding; during Ph.D. worked on the TRACE project and served as the supervisor for the ETH RoboCup team NomadZ; gained experience at Intel and interned at AWS; involved with the Square Kilometer Array project at Callaghan Innovation New Zealand, working on radio astronomy image reconstruction.
Education
- Ph.D. from ETH Zürich in 2024, supervised by Prof. Luc Van Gool and Dr. Martin Danelljan; M.Sc. in Advanced Signal Processing and Communications Engineering from FAU Erlangen-Nürnberg with highest distinction, Master's thesis at Johns Hopkins University under Prof. Mathias Unberath.
Background
- Research Interests: Robotic foundation models, 3D scene understanding. Professional Field: Efficient use of data in end-to-end algorithms for autonomous robotic systems, moving beyond costly robotic demonstrations.
Miscellany
- Contact: jan-nico.zaech@insait.ai; Address: INSAIT, Synergy Tower, 1784 Sofia, Bulgaria
