Papers: ICML 2025: ESPFormer: Doubly Stochastic Attention with Expected Sliced Transport Plans; ICLR 2025: ESP: Expected Sliced Transport Plans; ICLR 2025 (spotlight): LSSOT: Linear Spherical Sliced OT: A Fast Metric for Comparing Spherical Data; ICML 2024 (spotlight): S3W: Stereographic Spherical Sliced Wasserstein Distances.
Research Experience
Position: Machine Learning PhD Intern, Instacart, Fall 2025; Research Projects: ESPFormer, LOTFormer, LUNA, etc.
Education
PhD: Vanderbilt University, Computer Science, Advisor: Dr. Soheil Kolouri; B.Sc.: Amirkabir University of Technology (Tehran Polytechnic), Biomedical Engineering, Graduated in 2021.
Background
Research Interests: Transformers & LLMs, Optimal Transport, Applied Mathematics. Background: PhD candidate in Computer Science at Vanderbilt University, focusing on designing new attention mechanisms and uncertainty estimation methods for large language models.
Miscellany
Personal Interests: Playing Dota 2 (as a carry), exploring the world of coffee.