Research Assistant

SocialMind Lab, Researchbuddy AI
Feb 2025 – Present

• Evaluated reliability of large language models in low-resource, jurisdiction-specific legal domains.

• Designed evaluation pipelines combining automated metrics, LLM-as-judge scoring, and expert human assessment.

• Coordinated evaluations with licensed legal professionals and managed bias control protocols.

• Led manuscript drafting and dataset construction from authentic Bangladeshi legal forums.

• Applied statistical validation (ANOVA, Tukey’s HSD) with BLEU, ROUGE, and BERTScore.

Skills: Trustworthy AI, LLM Evaluation, Statistical Analysis, Research Writing

Research Apprentice

CUET NLP Lab, Chittagong University of Engineering & Technology
Sep 2024 – Jul 2025

A research lab specializing in Natural Language Processing, multilingual AI, and transformer-based models.

• Conducted research on multilingual text classification and low-resource NLP

• Worked with transformer models and large language models (LLMs)

• Contributed to research papers and ACL-affiliated workshops

• Supported experiments, analysis, and visualization for NLP research projects

Skills: Artificial Intelligence · Machine Learning · NLP · Large Language Models · Transformer Models · Python · LaTeX · Data Visualization

Publications and Ongoing Works

Ongoing Works
Title: Assessing the Reliability of Large Language Models in the Bengali Legal Context: A Comparative Evaluation Using LLM-as-Judge and Legal Experts

Authors: Sabik Aftahee, A.F.M. Farhad, Arpita Mallik, Ratnajit Dhar, Jawadul Karim, Nahiyan Bin Noor, Ishmam Ahmed Solaiman
Status: Journal of Experts system with application  (In Review, IF: 7.5)
Link: https://arxiv.org/abs/2511.05627

Title: rustworthy Explainable Artificial Intelligence in Oncology: Security, Fairness, and Future Directions

Status:
In Preparation (Review Paper)

Description:
A comprehensive review of explainable artificial intelligence (XAI) that critically examines the limitations of current explanation methods from the perspectives of security, explainability, fairness, robustness, and clinical validation. The paper proposes a taxonomy and evaluation framework for assessing trustworthy XAI in medical imaging by synthesizing existing evaluation metrics, identifying research gaps, and outlining future research directions toward clinically reliable AI systems.

Published Works

Title: MRAViT-XAI: A Novel Multi-Resolution Attention Vision Transformer Framework with Explainable AI for Enhanced Lung and Colon Cancer Classification

Authors: Md. Abdur Rahman, Sabik Aftahee, Lamim Zakir Pronay, Md. Ashiqur Rahman
Publication Venue: 2025 IEEE International Conference on Quantum Photonics, Artificial Intelligence, and Networking (QPAIN 2025)
IEEE Link: https://ieeexplore.ieee.org/document/11172105

Title: FAKD-XAI: Feature-Aligned Knowledge Distillation with Explainable AI for Efficient Brain Tumor Classification

Authors: Md. Abdur Rahman, Sabik Aftahee, Md. Ashiqur Rahman, Md. Pronay Farab
Publication Venue: International Conference on Data Science, AI and Applications (ICDSAIA 2025)
Springer Nature Link: https://link.springer.com/chapter/10.1007/978-3-032-11335-1_30

ACL Workshops:

Title: SmolLab_SEU at BEA 2025 Shared Task: A Transformer-Based Framework for Multi-Track Pedagogical Evaluation of AI-Powered Tutors

Authors: Md. Abdur Rahman, Md. Al Amin, Sabik Aftahee, Muhammad Junayed, Md. Ashiqur Rahman
Publication Venue: Proceedings of the 20th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2025)
Link: https://aclanthology.org/2025.bea-1.88/

Title: Zero_Shot at SemEval-2025 Task 11: Fine-Tuning Deep Learning and Transformer-Based Models for Emotion Detection in Multi-Label Classification, Intensity Estimation, and Cross-Lingual Adaptation

Authors: Ashraful Islam Paran, Sabik Aftahee, Md. Refaj Hossan, Jawad Hossain, Mohammed Moshiul Hoque
Publication Venue: Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)
Link: https://aclanthology.org/2025.semeval-1.247/

Title: Hinterwelt@LT-EDI 2025: A Transformer-Based Detection of Caste and Migration Hate Speech in Tamil Social Media

Authors: Md. Al Amin, Sabik Aftahee, Md. Abdur Rahman, Md. Sajid Hossain Khan, Md. Ashiqur Rahman
Publication Venue: Fifth Workshop on Language Technology for Equality, Diversity, and Inclusion (LT-EDI-2025), LDK 2025
Link: https://aclanthology.org/2025.ltedi-1.24/

Title: Hinterwelt@LT-EDI 2025: A Transformer-Based Approach for Identifying Racial Hoaxes in Code-Mixed Hindi-English Social Media Narratives

Authors: Md. Abdur Rahman, Md. Al Amin, Sabik Aftahee, Md. Ashiqur Rahman
Publication Venue: Fifth Workshop on Language Technology for Equality, Diversity, and Inclusion (LT-EDI-2025), LDK 2025
Link: https://aclanthology.org/2025.ltedi-1.21/

Title: CUET_NetworkSociety@DravidianLangTech 2025: A Transformer-Based Approach for Detecting AI-Generated Product Reviews in Low-Resource Dravidian Languages

Authors: Sabik Aftahee, Tofayel Ahmmed Babu, Md. Musa Kalimullah Ratul, Jawad Hossain, Mohammed Moshiul Hoque
Publication Venue: Proceedings of the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages
Link: https://aclanthology.org/2025.dravidianlangtech-1.110/

Title: CUET_NetworkSociety@DravidianLangTech 2025: A Multimodal Framework to Detect Misogyny Meme in Dravidian Languages

Authors: Md. Musa Kalimullah Ratul, Sabik Aftahee, Tofayel Ahmmed Babu, Jawad Hossain, Mohammed Moshiul Hoque
Publication Venue: Proceedings of the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages
Link: https://aclanthology.org/2025.dravidianlangtech-1.92/

Title: CUET_NetworkSociety@DravidianLangTech 2025: A Transformer-Driven Approach to Political Sentiment Analysis of Tamil X (Twitter) Comments

Authors: Tofayel Ahmmed Babu, Md. Musa Kalimullah Ratul, Sabik Aftahee, Jawad Hossain, Mohammed Moshiul Hoque
Publication Venue: Proceedings of the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages
Link: https://aclanthology.org/2025.dravidianlangtech-1.93/

Undergraduate Thesis:

Title: A Transformer-Based Approach for Identification of Natural Disasters from Multimodal Contents with Multimodal Fusion

Authors: Sabik Aftahee, Md. Monjur Ul Hasan
Status: Ongoing Journal Work
Manuscript: https://drive.google.com/file/d/15sE3Mcfk-4P5pULJ65l3hw1-kkG7Zxz7/view?usp=sharing

Preprints and Other Research Activities