Mohammadamin Rami

Hi there! I am Mohammadamin “Amin” Rami, scientific researcher at ISTA, where I work in the Data Science, Machine Learning, and Information Theory group under the supervision of Prof. Marco Mondelli. My research focuses broadly on the mathematical foundations of AI, at the intersection of machine learning, high-dimensional statistics, and information theory.

I received my Master’s degree from Institut Polytechnique de Paris, where I was fortunate to be advised by Prof. Aslan Tchamkerten. Prior to that, I earned my Bachelor’s degree in Electrical Engineering from Sharif University of Technology.

news

Sep 2026	Our paper “Effective Context in Transformers: An Analysis of Fragmentation and Tokenization” has been accepted to Neurips 2026! The paper will have a revised title: “Representation in Finite-Context Prediction: An Analysis of Tokenization and Fragmentation”.
Sep 2026	I have completed my MSc at Institut Polytechnique de Paris!

selected publications

Effective Context in Transformers: An Analysis of Fragmentation and Tokenization

Amirmehdi Jafari Fesharaki, Mohammadamin Rami, and Aslan Tchamkerten

May 2026

Abs arXiv Bib HTML

Transformers predict over a representation of a sequence. The same data can be written as bytes, characters, or subword tokens, and these representations may be lossless. Yet, under a fixed context window, they need not expose the same information to the model. This raises a basic question: how does the choice of representation change what a finite-context predictor can achieve?
@article{rami2026effectivecontext, title = {Effective Context in Transformers: An Analysis of Fragmentation and Tokenization}, author = {Jafari Fesharaki, Amirmehdi and Rami, Mohammadamin and Tchamkerten, Aslan}, year = {2026}, month = may, url = {https://arxiv.org/abs/2605.13485}, }