I am a PhD student at the Visualization Research Center (VISUS) of the University of Stuttgart, advised by Prof. Dieter Schmalstieg, and a member of the International Max Planck Research School for Intelligent Systems (IMPRS-IS). My research focuses on open-vocabulary 3D scene understanding, teaching machines to recognize, organize, and reason about the objects and relationships that make up 3D environments.
I am especially interested in 3D scene graphs as a structured representation for embodied AI. Besides my main research focus, I worked on egocentric action recognition during my masters and still pursue side projects in this direction.
Before my PhD, I worked alongside my studies in areas ranging from consulting and enterprise architecture to data engineering, including roles at SAP, Deloitte, and Mercedes-Benz.
An asynchronous architecture that decouples lightweight online mapping from heavyweight VLM reasoning, keeping the 3D scene graph queryable from the first frame while background agents progressively enrich it.
University of Stuttgart 路 VISUS
Open-vocabulary 3D scene understanding and semantic scene graph modeling in egocentric environments. Advised by Prof. Dieter Schmalstieg.
University of Stuttgart
Thesis: Multi-modal Graphormer for Action Recognition in Egocentric Videos (1.0), a hierarchical multi-stream model combining a Graphormer (GNN) with Vision Transformers over hand-object interactions, motion, gaze, and scene semantics.
Esslingen University of Applied Sciences
Thesis: Monocular 3D Stixel-Net, a CNN for object detection and depth estimation in traffic scenes using LiDAR and RGB from the Waymo Open Dataset.