Damiano Marsili

I'm a Ph.D student at Caltech advised by Georgia Gkioxari and Pietro Perona. Before my Ph.D, I was an Applied Science Intern at Amazon Robotics working on 3D spatial reasoning. Prior to that, I double-majored at Johns Hopkins in Computer Science and Mathematics.

I'm interested in pushing the reasoning capabilities of multimodal LLMs. Most of my research involves post-training vision-language models for visual reasoning.

I'm best reachable via email at dmarsili at caltech dot edu.

Email / Resume / Google Scholar / Twitter/X / LinkedIn / Github

News

[Feb. 2026] TWIN was accepted to CVPR 2026!
[Jan. 2026] VALOR was accepted to ICLR 2026!
[Feb. 2025] VADAR was accepted to CVPR 2025!
[Sep. 2023] Started my Ph.D at Caltech working with Georgia Gkioxari and Pietro Perona.
[Jun. 2023] Started an internship at Amazon Robotics working on 3D Spatial Reasoning
[May 2023] Graduated with a double major in Computer Science and Mathematics.
[Aug. 2020] Started my undegraduate at Johns Hopkins.

Publications

	TWIN - Same or Not? Enhancing Visual Perception in Vision-Language Models Damiano Marsili, Aditya Mehta, Ryan Lin, Georgia Gkioxari CVPR, 2026 project page / arXiv / code / bibtex
	VALOR - No Labels, No Problem: Training Visual Reasoners with Multimodal Verifiers Damiano Marsili, Georgia Gkioxari ICLR, 2026 project page / arXiv / code / bibtex
	VADAR - Visual Agentic AI for Spatial Reasoning with a Dynamic API Damiano Marsili, Rohun Agrawal, Yisong Yue, Georgia Gkioxari CVPR, 2025 project page / arXiv / code / bibtex

Website Template credits: Jon Barron