About

I work at the boundary between neuroscience and computer science, exploring questions in NeuroAI that trace back to the cybernetics tradition. I'm particularly interested in understanding how neural networks develop internal representations to solve cognitive tasks.

I am currently a medical doctor, having completed my education at Ahvaz University of Medical Sciences. These are my final days in the medical profession before I make a total career shift.

Google Scholar / Github / Twitter / Email

Research Interests

Reinforcement Learning World Models Cognitive Maps NeuroAI

Experience

Research Assistant

CMC Lab, TU Dresden

June 2024 - May 2025

Trained RNNs on a set of cognitive tasks from the NeuroGym library and used dynamical similarity analysis (DSA) to investigate different factors affecting dynamics. Research focuses on understanding how task optimization and input characteristics influence neural network representations.

Education

Doctorate of Medicine

Ahvaz University of Medical Sciences

2017 - 2025

GPA: 3.6/4.0, Top 20%. Graduated from one of the top medical universities in the country. Medicine is a 7-8 year program in Iran and in its last year we work as medical interns before graduating with an M.D. degree.

Conference Abstracts

Heterogeneous Effect of Input and Task-optimization on the Dynamics of RNNs

CCN 2025

(Not presented due to visa issues)

PDF

Identifying task-specific dynamics in recurrent neural networks using Dynamical Similarity Analysis

Bernstein 2024

Skills

Programming

Python, C, C++, Bash

Machine Learning Toolkits

Pandas, NumPy, Gymnasium, TorchRL, HuggingFace, VLLM

Operating Systems

Linux, Windows

Contact

Feel free to reach out if you're interested in collaboration or have questions about my research. You can find my contact information and links in the About section above.