Hi there 👋, I'm Masoud, a graduate student in Computer Science at the University of Alberta, supervised by Prof. Osmar R. Zaïane, and currently a Research Intern at Electronic Arts (EA) 🎮.
My thesis research focuses on visual and spatial reasoning in multimodal language models. Currently, I am exploring how inference-time scaling, adaptive context control, and RL post-training can improve the efficiency and adaptability of spatial reasoning. Spatial reasoning has broad applications across domains including robotics, autonomous driving, video games, and VR/AR.
At EA, my research focuses on developing small, efficient language models for real-time decision-making and video game applications.
- Efficient Spatial & Visual Reasoning with LLMs/VLMs/MMLMs
- Vision-Language Understanding & Embodied Spatial Reasoning
- 3D Representations, Grounding, & Space Understanding
- Building Vision-Language Datasets for Post-training MLLMs on Spatial Reasoning Tasks
- Visual and Geometry Retrieval Systems
- M.Sc. in CS, University of Alberta (Present)
- Ph.D. in ECE, University of Alberta (Transferred to CS)
- M.Sc. & B.Sc. in ME, Sharif University of Technology & Univ. of Tehran
