Md. Ataullha

Work History

• AI Engineer
  Synesis IT PLC, Dhaka, Bangladesh
  Sep 2024 – Mar 2025
  Low Resource Chatbot Development
• NLP Engineer
  Synesis IT PLC, Dhaka, Bangladesh
  Mar 2024 – Aug 2024
  Low Resource ASR, and Callbot Development

Extra Curricular Activities

• Reviewed papers for ICCIT 2024.
• Mentoring 2+ undergraduate thesis teams from SUST during undergraduate on: Bangladeshi Regional Audio Classification, Ethnic Language-Based Text-to-Speech (TTS) Synthesis System, and Dataset Creation with Prof. Dr. M. Shahidur Rahman (Sir).
• Mentoring 3+ undergraduate thesis teams from SUST on: LLMs Evaluation, Regional Language Translation, and Dataset Creation with Prof. Dr. M. Shahidur Rahman (Sir).
• Co-hosted DL Enigma 1.0, part of SUST CSE CARNIVAL 2024.

Teaching

• Planned to teach CSE 435, CSE 436, CSE 201, CSE 202 at Fall, 2025
• CSE 411: Machine Learning (223_D2) [Spring, 2026]
• CSE 412: Machine Learning Lab (223_D1, 223_D3) [Spring, 2026]
• CSE 436: Data Mining Lab (221_D2) [Fall, 2025]
• CSE 436: Data Mining (221_D11) [Fall, 2025]
• CSE 201: Object Oriented Programming Lab (242_D2) [Fall, 2025]
• CSE 202: Object Oriented Programming Lab (242_D4) [Fall, 2025]
• CSE 411: Machine Learning (221_D10) [Summer, 2025]
• CSE 412: Machine Learning Lab (221_D20) [Summer, 2025]
• CSE 436: Data Mining Lab (221_E2) [Summer, 2025] (213_D5, 213_E1) [Spring, 2025]
• CSE 100: Computational Thinking and Problem Solving (250_D4) [Summer, 2025]
• CSE 435: Data Mining (212_D3) [Spring, 2025]
• CSE 406: Integrated Design Project II (221_D15) [Spring, 2025]
• CSE 310: Operating System Lab (223_E2) [Spring, 2025]
• CSE 202: Object Oriented Programming Lab (241_D5) [Spring, 2025]

Thesis Supervision / Co-Supervision

• Supervising 1 thesis group parallely with Sadia Afroze working on Deepfake Speech [Summer, 2025] (No further supervision or co-supervision commitments at GUB planned unless recommend by a Professor or SUST alumni faculty).
• Co-supervising 2+ thesis groups working on Machine Translation based LLM Evaluation with Prof. Dr. Md Saiful Azad (Sir) [Summer, 2025].
• Co-supervising 1+ thesis groups working on Natural Language Processing with Sadia Afroze [Summer, 2025].
• Co-supervising 1+ project groups working on a Medical Image Diagnosis System with Md. Riad Hassan [Spring, 2025].
• Co-supervising 2+ thesis groups working on Speech Processing and Reinforcement Learning with Dr. Muhammad Aminur Rahaman [Fall, 2024].

Google Scholar Profile: https://scholar.google.com/citations?user=N2dzzwsAAAAJ&hl=en

Journal Publications

1. Md Shahidul Islam, Pabel Shahrear, Goutam Saha, Md Ataullha, and M. Shahidur Rahman.
   Mathematical Analysis and Prediction of Future Outbreak of Dengue on Time-Varying Contact Rates Using a Machine Learning Approach.
   Computers in Biology and Medicine, Vol. 178, 2024, Article 108707.
   DOI: https://doi.org/10.1016/j.compbiomed.2024.108707

Conference Proceedings

1. Md Ataullha, Mushfiqur Rahman, and M. Shahidur Rahman.
   PakhiderChobi: A Comprehensive Dataset for Real-Time Detection of Bangladeshi Birds.
   In Proceedings of the 26th International Conference on Computer and Information Technology (ICCIT), IEEE, 2023.

2. Md Ataullha, Soumik Paul Jisun, Ishrat Jahan, and Mushfiqur Rahman.
   BanFish: A Dataset for Classifying Common Bangladeshi Fish Species.
   In Proceedings of the 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT), IEEE, 2024.
   DOI: https://doi.org/10.1109/ICEEICT62016.2024.10534462

3. Md Ataullha, Mahedi Hassan Rabby, Mushfiqur Rahman, and Tahsina Bintay Azam.
   Bengali Document Layout Analysis with Detectron2.
   arXiv preprint, arXiv:2308.13769, 2023.
   https://arxiv.org/abs/2308.13769

4. Md. Hasibul Hasan, Sourav Podder, Syed Abdullah, Saiful Azad, Saurav Chandra Das, and Md Ataullha.
   D-D3N: Dropout-Regularized Deep Neural Networks for Android Malware Detection.
   Department of CSE, Green University of Bangladesh; Gazipur Digital University.
   (4th IEEE International Conference on Robotics, Automation, Artificial-intelligence and Internet-of-Things)

5. Md Faisal Hasan, Mst Rokshanara Toma, Md Ataullha, Sharifur Rahman, and M. Shahidur Rahman.
   CXR-Next: An Explainable Multi-Class Deep Learning Framework for Thoracic Disease Classification from Chest X-Rays.
   Department of CSE, Green University of Bangladesh; Department of CSE, SUST.
   (International Conference on Intelligent Data Analysis and Applications (IDAA 2025))

6. Md. Muzahidul Islam, Md. Ataullha, Sharifur Rahman, and M. Shahidur Rahman.
   Enhancing Code Synthesis with Bangla Instructions via Fine-Tuned Gemma-3.
   Department of CSE, Green University of Bangladesh; Department of CSE, SUST.
   (International Conference on Intelligent Data Analysis and Applications (IDAA 2025))

7. Md. Naimul Haque, Mostahid Hasan Fahim, M. Reza Selim, Md. Minhajul Haque, Md. Fuad Al Amin, Md. Ataullha, and M. Shahidur Rahman.
   Evaluating and Enhancing Large Language Models for Islamic Question Answering.
   Department of CSE, SUST; Department of CSE, Green University of Bangladesh.
   (28th International Conference on Computer and Information Technology)

8. Jahidul Islam, Md Ataullha, and Saiful Azad.
   PyBangla at BLP-2025 Task 2: Enhancing Bangla-to-Python Code Generation with Iterative Self-Correction and Multilingual Agents.
   Department of CSE, Green University of Bangladesh.
   (International Joint Conference on Natural Language Processing & Asia-Pacific Chapter of the Association for Computational Linguistics, 2025)

9. Adnan Sami Pavel, Sraboni Akhter, Asha Mony, Md. Ataullha, and Muhammad Aminur Rahaman.
   Structured Prompting and a Heterogeneous LLM Multi-Agent Framework for Financial Sentiment Analysis.
   Department of CSE, Green University of Bangladesh; Department of CSE, BUBT.
   (IEEE i-COSTE 2025 SUSTAINABLE TECHNOLOGY FOR HUMANITY)

• Kaggle Notebook Expert (1x Expert) [Kaggle Profile]

My research explores innovative applications of Artificial Intelligence in low-resource environments, with a strong emphasis on:

• Large Language Models (LLMs)
• Bangla Speaker Diarization
• Bangla Automatic Speech Recognition (ASR)
• Bangla Text-to-Speech Synthesis (TTS)
• Machine-learning-driven AI products
• Domain Specific Chatbot
• Time Series Forecasting
• Object Detection
• Dataset Creation

I aim to advance speech processing, natural language understanding (NLU), and text-to-speech (TTS) systems to develop accessible AI products that drive progress in technology, education, and conservation.

Supervisor

• Dr. Mohammad Shahidur Rahman
  Professor
  Department of Computer Science and Engineering
  Shahjalal University of Science and Technology, Sylhet, Bangladesh
  Email: rahmanms@sust.edu
  Mobile: +880 1914-930807
  Faculty Profile: https://www.sust.edu/departments/cse/faculty/rahmanms@sust.edu

Relation: Thesis Supervisor & Mentor
Project Involvement: Actively engaged in 2+ research projects under his guidance.

Featured Projects

• StarGaze BD (ML Algorithms: RF, SVM, LR, CNN, Flask, JavaScript, Streamlit)
  A web application for identifying prominent Bangladeshi personalities from photos using machine learning and deep learning. It leverages custom ML algorithms (Random Forest, SVM, Logistic Regression) and a CNN model for high accuracy, with Flask handling backend processing and a user-friendly interface built with JavaScript and Streamlit. [GitHub Link]
• Freshman Utilities (Java, Firebase, Google Directions API, Jsoup)
  An Android app designed to assist university freshmen with campus navigation using Google Directions API, Firebase-based resource sharing, and web scraping with Jsoup to collect essential campus-related information. [GitHub Link]
• CourseTrackr (Java, Servlet, JSP, MySQL)
  A web-based course management system featuring role-based access for admins, teachers, and students. Developed using Java Servlets and JSP for dynamic web pages and MySQL for backend database management. [GitHub Link]
• QuesTek Pro (Java, JavaFX, Text Files)
  A JavaFX-based desktop application for conducting and managing multiple-choice quizzes. It includes admin verification, quiz result exportation, and local data storage using text files—ideal for in-person examinations. [GitHub Link]

Recent Talks & Interviews

• Presentation at ICCIT’23 (26th ICCIT, Cox’s Bazar, December 2023)
  Delivered a presentation on PakhiderChobi, a dataset featuring 8,670 annotated images of 33 Bangladeshi bird species. The talk highlighted the importance of accurate bird detection and classification for urban conservation, leveraging the YOLOv8 object detection model to achieve 95.3% mAP with an inference time of 6.6 milliseconds per image. This tool supports real-time bird recognition, even in challenging environments, contributing to conservation efforts in the face of urbanization and habitat loss. [Watch on YouTube]