Francis Engelmann

Francis Engelmann is an incoming Assistant Professor at the University of Lugano (USI) Faculty of Informatics, currently completing his postdoctoral research at Stanford University working with Prof. Leonidas Guibas and Prof. Jeannette Bohg. Prior to Stanford, he was a postdoctoral research fellow at ETH Zurich with Prof. Dr. Marc Pollefeys and conducted his PhD in Computer Vision, Machine Learning and 3D Scene Understanding at RWTH Aachen University under Prof. Dr. Bastian Leibe.

His research focuses on advancing 3D scene understanding through computer vision, machine learning, and robotics. Engelmann's work explores open-vocabulary 3D scene understanding, functional scene analysis, 3D reconstruction, and semantic segmentation. He has made significant contributions to open-vocabulary 3D instance segmentation, 3D scene graph construction, and language-augmented 3D vision. His research bridges theoretical computer vision with practical robotics applications, particularly in scene understanding for robotic manipulation.

Analysis of his recent publications reveals a strong trajectory in developing methods for open-vocabulary 3D scene understanding, with increasing focus on functional understanding, language integration, and practical robotics applications. His work spans from foundational 3D representation learning to applied robotics systems, with a consistent emphasis on making 3D scene understanding more accessible, scalable, and semantically rich.

• ETH Zurich Career Seed Award (2022)
• Outstanding Reviewer for CVPR 2024 (top 2%)
• Top Reviewer for NeurIPS 2023
• Best Paper Award at ICRA'24 MOMA.v2 workshop
• Multiple oral presentations at top-tier conferences (ICCV, CVPR, ECCV)

Engelmann actively mentors PhD students and postdocs, with notable collaborations including Valentin Bieri, Rui Huang, Elisabetta Fedele, and Ayca Takmaz. He has received an NVIDIA Academic grant and serves as an area chair for major computer vision conferences including CVPR'25, WACV'26, and 3DV'25. He co-organizes the annual Open-Vocabulary 3D Scene Understanding Workshop, now in its fourth iteration at CVPR'25.