S a f i n a l

Hello! Iโ€™m

Ali Nafisi

  • mail
  • googleScholar
  • X
  • LinkedIn
  • github
  • cv

Iโ€™m currently completing my MS in AI at Bu-Ali Sina University. I earned my BS in Computer Engineering from the University of Kurdistan in 2023.

My research experience is rooted in Deep Learning, specifically Natural Language Processing, Multimodal Models, and Trustworthiness. I am now aiming to expand my research scope to explore Neuro-Symbolic AI and Reinforcement Learning.

My current work explores the use of diffusion models for counterfactual data augmentation to mitigate spurious correlations in image datasets.

Iโ€™m actively seeking a PhD position where I can continue to grow as a researcher and contribute to cutting-edge, high-impact projects.

News

November 2025

Ranked ๐Ÿฅ‡ 1st in the Rayan International AI Contest! The technical report of our solutions is available on arXiv.

May 2025

Ranked 3rd in the SUT LLM Agents Hackathon! ๐Ÿ†

May 2025

Our paper on Culturally Aligning LLMs has been accepted to ACL! ๐ŸŽ‰

March 2025

I joined Robot Intelligence & Vision Lab as a Graduate Student Researcher @ Bu-Ali Sina University.

Publications

alt
SB. Hosseinbeigi | MA. SeifKashani | J. Seraj | F. Taherinezhad | A. Nafisi | F. Nadi | I. Barati | H. Hasani | M. Amiri | M. Masoudi
alt
F. Daneshfar | S. Soleymanbaigi | A. Nafisi | P. Yamini

Experience

Graduate Research Assistant

Robot Intelligence & Vision Lab

Work Study

March 2025 - Present

Exploring the use of diffusion models for counterfactual augmentation to mitigate spurious correlations in image datasets.

AI Engineer

Matina AI

Part-time

April 2024 - January 2025

Investigated and implemented culturally-aligned LLM post-training approaches by creating specialized datasets and leveraging Multiple LoRA Experts, resulting in the model Matina and a published paper at ACL 2025.

Course Instructor

Quera

Part-time

March 2021 - February 2022

Developed and instructed programming courses, including creating tutorials, designing unit-tested problems, and providing debugging support to students.