Ariyan Hossain

About Me

📢 Looking for Ph.D. opportunities for Fall 2026 / Spring 2027

I'm Ariyan Hossain, a Lecturer in the Department of Computer Science and Engineering at BRAC University, Dhaka, Bangladesh.

My research interests span trustworthy and responsible AI, along with security and privacy in language models. Particularly, I am interested in developing methods to identify and reduce bias in AI systems, promoting fairness and inclusion across communities. I also aim to design trustworthy and transparent LLMs that are robust, interpretable, and privacy-preserving to benefit society.

I graduated with Highest Distinction (GPA: 4.0/4.0) from the Department of Computer Science and Engineering, BRAC University, where I completed my undergraduate thesis on gender bias mitigation in transformer models under the supervision of Dr. Farig Yousuf Sadeque.

My Resume

Work Experience

Lecturer

BRAC University

Department of Computer Science and Engineering

Dhaka, Bangladesh

May 2024 — Present

Teaching CSE220 (Data Structures), CSE221 (Algorithms), and CSE440 (Natural Language Processing)
Taking both theory and lab classes, creating lecture notes, invigilating exams, and grading scripts
Supervising thesis students and serving as thesis defense panel judge

Research Assistant

BRAC University

Department of Computer Science and Engineering

Dhaka, Bangladesh

November 2023 — May 2024

Worked full-time under Dr. Muhammad Nur Yanhaona
Classified framing of news articles using Machine Learning and NLP
Attended weekly research meetings and maintained detailed documentation

Undergraduate Teaching Assistant

BRAC University

Department of Computer Science and Engineering

Dhaka, Bangladesh

October 2022 — August 2023

Assisted in labs of CSE221 (Algorithms)
Graded assignments and gave occasional lectures
Provided up to 15 hours of weekly consultation to undergraduate students

Research Experience

Exploring Bengali Dialectal Bias and Mitigation Strategies in Multilingual Large Language Models

K. M. J. Sami, D. Sumit, Ariyan Hossain, Dr. Farig Yousuf Sadeque

BLP Workshop @ IJCNLP-AACL 2025 (Accepted) (2025)

Developing a two-phase framework integrating Retrieval-Augmented Generation (RAG) and Reinforcement Learning from AI Feedback (RLAIF) to detect and quantify dialectal bias across four major Bengali dialects. Designing two RAG pipelines for standard-to-dialectal translation and constructing a gold-labeled benchmark dataset of 300 standardized questions per dialect.

ACL Anthology

Exploring and Mitigating Gender Bias in Transformer-based Language Models

Ariyan Hossain, K. M. A. Hannan, R. Haque, N. T. Rafa, H. Musarrat, S. A. Dipu, Dr. Farig Yousuf Sadeque

Undergraduate Thesis (2025)

Proposed MALoR, a novel metric for quantifying gender bias via masked-token probabilities in encoder-based models. Created gender-balanced pretraining datasets using Counterfactual Data Augmentation (CDA) by swapping gendered pronouns and names, generating over 19,000 balanced sentence pairs. Achieved up to 95% reduction in gender bias (e.g., "he–she" bias in BERT-base decreased from 1.27 → 0.08 and "his–her" from 2.51 → 0.36 MALoR score).

arXiv

Identifying Framing Bias in Online Climate Change News Articles

Ariyan Hossain, Dr. Muhammad Nur Yanhaona

Building a frame-classified climate change news dataset using expert annotation and synthetic data augmentation through LLM (GPT-4) prompt engineering. Performing multi-class framing classification with Longformer, BigBird, BERT, and Llama-3.1 architectures for long-text handling. Applying Parameter-Efficient Fine-Tuning (PEFT) QLoRA to finetune LLaMA to reduce memory requirements.

Education

Bachelor of Science in Computer Science and Engineering

BRAC University

Dhaka, Bangladesh

January 2020 — November 2023

GPA: 4.0/4.0; Graduated with Highest Distinction
Relevant Coursework: Natural Language Processing, Neural Networks, Artificial Intelligence, Linear Algebra, Calculus