Ph.D. Candidate in Computer Science, University of Vermont
I am a Ph.D. candidate in Computer Science at the University of Vermont, advised by Prof. Safwan Wshah. My research is in computer vision, focused on visual geo-localization: finding where an image was taken by matching it to satellite imagery, from large-scale cross-view retrieval to real-time, fine-grained pose estimation. I also work on localization from RF signals with calibrated uncertainty, and on vision-language reasoning for spatial understanding.
Before my Ph.D., I spent four years at Orange Jordan as an RF engineer and senior data analyst. In summer 2026 I interned at NextNav, building probabilistic machine learning models for 5G positioning. I hold an M.S. in Computer Science from the University of Vermont and a B.S. in Electrical Engineering from the University of Jordan.
I am on the job market for research scientist and applied scientist roles starting fall 2027.
My work is aimed at spatial intelligence: giving machines a reliable sense of where they are and how the scene around them is laid out. Within that, I am interested in cross-view image geo-localization, 3D vision (reconstruction, feature matching, pose estimation), generative models (flow matching, diffusion), vision-language models, and multimodal localization.
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026
[paper] [arXiv] [code] [project]
Real-time, fine-grained ground-to-satellite localization that refines many pose hypotheses through flow-inspired iterative regression, with control over the speed and accuracy trade-off at test time.
NeurIPS 2026 Workshop on Physical World AI
A training-free vision-language pipeline that describes ground and satellite views as structured text for cross-view matching, showing that scene structure survives the move into language while the fine appearance detail needed to separate nearby places does not.
IEEE International Conference on Communications (ICC), 2026
Spatially adaptive conformal prediction inside a graph transformer, giving location-aware uncertainty regions with statistical coverage guarantees for Wi-Fi indoor localization.
IEEE Internet of Things Journal, 2025
A unified framework for anchor-free and anchor-based sensor localization from RSSI alone, modeling how signals change over time and how nodes relate in space.
International ISCRAM Conference, 2024