Welcome to my homepage!

I am a researcher working on the mathematical foundations of Artificial Intelligence (AI) with the goal of developing the next generation of safe and trustworthy AI systems. Currently, I am working on the Universality Phenomenon and a Non-Probablistic Theory of Machine Learning.

My Ph.D. at the IDEAS Center of the University of Pennsylvania advised by Prof. Hamed Hassani focused on developing mathematical techniques to analyze tractable high-dimensional models of deep learning, and using them to better understand modern learning systems. I was fortunate to be able to collaborate extensively with Prof. Hamed Hassani, Prof. Edgar Dobriban, Dr. Donghwan Lee, Dr. Thomas T.C.K. Zhang, and Prof. Nikolai Matni. Before that, I obtained my M.A. in Statistics from the Wharton School of the University of Pennsylvania, and my B.Sc. in Electrical Engineering from Sharif University of Technology (with highest distinctions).

Research Interests

My interest is in Deep Learning {Theory, Optimization, Phenomenology}, Random Matrix Theory, High-Dimensional Statistics, Information Theory, and Quantitative Finance. My work addresses several questions:

• How do neural networks learn features? [ICML 2024] [ArXiv 2026]
• How to design deep learning optimizers that learn better features? [ICML 2025]
• Deep Learning Phenomenology:
  • Weak-to-Strong Generalization [NeurIPS 2025]
  • Agreement-on-the-Line [ICML 2023]
  • Double Descent with Correlated Samples [L4DC 2025]
• On the Information-Theoretic nature of excess risk: [IEEE Trans. Info. Theory] [ICML 2021]

A full publication list can be found here. My research is supported by NSF TILOS and a gift from Amazon AWS AI.

Work Experiences

• Machine Learning Researcher, Cubist Systematic Strategies (Point72), Currently.
• Machine Learning Research Intern, Cubist Systematic Strategies (Point72), Summer 2025.
• Summer Data Science Intern, Digikala, Summer 2019.
  

Teaching Experiences

Teaching is one of my favorite academic activities, and I have been the TA for many courses so far:

• ESE 6740: Information Theory, Prof. Shirin Saeedi Bidokhti.
• CIS 5200: Machine Learning, Prof. Surbhi Goel and Prof. Eric Wong.
• ESE 5420: Statistics for Data Science, Prof. Hamed Hassani. [Outstanding Teaching Award]
• I have also TA’ed for many course at Sharif; e.g., Theory of Machine Learning, Probability in High Dimensions, High Dimensional Statistics, Causal Inference, Probability and Statistics, Signals and Systems, and Engineering Mathematics.
  

Other Interests

Philosophy of Probability, History of Probabilistic Ideas, Causality and Causal Inference, Financial Markets, Quantitative Portfolio Management, 13th-15th Century Persian Poetry, Contemporary Persian Poetry, Politics, U.S. Immigration Law, Federal Civil Litigation, and Antiquarian Books.

I could fill Volumes with Descriptions of Temples and Palaces, Paintings, Sculptures, Tapestry, Porcelaine, &c. &c. &c. — if I could have time. But I could not do this without neglecting my duty. The Science of Government it is my Duty to study, more than all other Studies Sciences: the Art of Legislation and Administration and Negotiation, ought to take Place, indeed to exclude in a manner all other Arts. I must study Politicks and War that my sons may have liberty to study Painting and Poetry Mathematicks and Philosophy. My sons ought to study Mathematicks and Philosophy, Geography, natural History, Naval Architecture, navigation, Commerce and Agriculture, in order to give their Children a right to study Painting, Poetry, Musick, Architecture, Statuary, Tapestry and Porcelaine.

John Adams, May 12th, 1780.