Hello, World!

I am Zahra Sodagar, a Ph.D. student in Computer Science at the University of Maryland, advised by Prof. Soheil Feizi, and collaborating with Prof. Sarah Wiegreffe. I graduated with a B.Sc. in Electrical Engineering with a minor in Applied Mathematics, from Sharif University of Technology.

My research mostly focuses on studying the emergence of reasoning and planning behaviours in generative models, aiming to systematically improve them through scalable methods for adaptation, moving towards solving “Intelligence”.

I approach this from a principled perspective, grounded in statistics and information theory, while drawing on ideas from psychology and cognitive science to guide how we model and refine intelligent behavior. Rather than treating reasoning as a byproduct, I view it as a process that can be explicitly analyzed, improved, and engineered. More broadly, my goal is to move from scaling behavior to understanding and improving these reasoning processes, toward models that are interpretable, controllable, generalizable and dynamic, yet robust and reliable.

Explore More

About Me — Background, interests, and current directions.

Publications — Selected publications and recent projects.

Posts — Ideas, images, and things I find worth sharing often outside academia.