Danilo Đorđević

ETH Zürich, Doctoral Researcher

Danilo Dordevic is a doctoral researcher at ETH Zurich’s Artificial Visual Intelligence Group, supervised by Prof. Konrad Schindler and Dr. Christos Sakaridis. He works on vision-language models (VLMs) and vision-language-action models (VLAs) for embodied AI. His research pushes VLAs beyond flat 2D image tokens toward explicit 3D scene grounding, equips them to acquire information actively, including asking under uncertainty, and couples them with predictive world models for hierarchical planning. He aims to build embodied agents that ground language reliably and anticipate the consequences of their actions.