Alireza Farhadi

Alireza Farhadi

Hi, I am a computer science researcher. My recent publications have focused on Large Language Models, Generative AI, and Multi-Agent Systems. However, my expertise and interests are not limited to these areas. I am open to collaborations across broader fields of Artificial Intelligence and Data Science, and I believe that with my experience and an interdisciplinary perspective, I can contribute to addressing diverse and complex computational challenges in new and innovative ways.

Education

AUT

Master of Science in Computer Science

Amirkabir University of Technology (Tehran Polytechnic)

2024 – Present
IUT

Bachelor of Mathematical Sciences and Applications

Minor in Computer Engineering, Isfahan University of Technology

2019 – 2024
NODET

Diploma in Mathematics and Physics

Allameh Helli High School, National Organization for Development of Exceptional Talents

2015 – 2019

Publications

Brains vs Bytes

Brains vs. Bytes: Evaluating LLM Proficiency in Olympiad Mathematics

Conference on Language Modeling (COLM 2025)
Reposted on X by Prof. Yann LeCun
PDF
RefGrader

RefGrader: Automated Grading of Mathematical Proofs

Under review at ICLR 2026
PDF
CombiGraph

CombiGraph-Vis: A Multimodal Olympiad Benchmark

Under review at NeurIPS 2025 Workshop
PDF
IteraGraph

IteraGraph: Multi-LLM Iterative Answering and Verification on Graphs

Working paper

News

Oct 25
Reviewed two papers for NeurIPS Workshop MATH-AI 2025.
Jul 25
Our paper "Brains vs. Bytes: Evaluating LLM Proficiency in Olympiad Mathematics" is accepted to COLM 2025.

Research Experience

Sharif

Graduate Research Assistant

Sharif University of Technology (ISDS Center)

I design and implement scalable multi-agent reinforcement learning frameworks aimed at enabling coordination, communication, and adaptive decision-making among intelligent agents in complex dynamic environments.

2025 – present
SigmaPulse

Founder & AI Research Engineer

SigmaPulse

I founded a startup focused on applying the latest advancements in artificial intelligence — including generative and multi-agent models — to financial data analysis, enabling more interpretable, adaptive, and context-aware predictions of market behavior.

2023 – present

Honors & Awards

🎓

Ranked 4th by GPA

Among undergraduate students, Department of Mathematical Sciences, Isfahan University of Technology

2020 – 2024
🧠

Ranked 6th

ACM International Collegiate Programming Contest (ICPC), Tehran Regional

2021 – 2022
🥇

Gold Medal

20th Bulgaria International Mathematics Competition (BIMC), Iran Regional

2013 – 2014
🥇

Gold Medal

2nd Mathematical Olympiad of Iranian Teenagers

2012 – 2013