About me

Hi! I’m Pouya Sadeghi, a recent Computer Engineering graduate with a passion for making AI systems more reliable, safe, and inclusive.

My academic journey has centered around building intelligent systems that are not only powerful but also reliable and fair. I work at the intersection of large language models, multilingual NLP, and AI safety, with a strong interest in both theoretical insights and practical impact.

Looking ahead, I’m eager to build on this foundation by creating AI tools that not only push the boundaries of what machines can do, but also earn people’s confidence through transparency, reliability, and integrity. I’m driven by the goal of shaping technology that truly serves society.

News

• 2025.04: Our paper, “Supermartingale Certificates for Quantitative Omega-regular Verification and Control,” has been accepted to CAV 2025.

• 2024.12: Our paper, “Extending LLMs to New Languages: A Case Study of LLaMA and Persian Adaptation,” has been accepted to COLING 2025.

• 2024.08: I joined Fatima Fellow as a fellow, working on Multilingual Multicultural Red Teaming of LLMs.

• 2024.08: I joined Singapore Management University (SMU) as a Visiting Research Student, working on Formal Methods for Safe and Trustworthy Probabilistic Systems.

• 2024.03: Awarded the Gold Medal for achieving 1st place at the 28th Scientific Olympiad of University Students of Iran (Computer Science branch).

• 2024.03: Our paper, “uTeBC-NLP at SemEval-2024 Task 9: Can LLMs be Lateral Thinkers?,” along with two other papers, has been accepted to SemEval@NAACL 2024.

• 2024.02: Our paper, “Benchmarking Large Language Models for Persian: A Preliminary Study Focusing on ChatGPT,” has been accepted to LREC-CoLing 2024.

• 2022.09: I joined the University of Tehran to pursue my B.Sc. in Computer Engineering.