About
I am a Lecturer in Computer Science and Engineering at Southeast University, Dhaka (since October 2022), where I teach core undergraduate courses and supervise student research. My work spans three connected areas: building datasets and benchmarks for low-resource Bangla NLP — covering regional dialects, named entities, and hate speech; applying transformer architectures to time-series forecasting in financial markets; and developing explainable deep learning methods for medical and agricultural imaging.
I graduated from Ahsanullah University of Science and Technology (AUST) in 2022 with a CGPA of 3.77/4.00, ranked 3rd of 139 graduates (Top 3%, Dean's List of Honour). My research outputs include 22 publications (15 peer-reviewed), four open datasets hosted on Mendeley Data, and reproducible code on GitHub. I serve as Co-Investigator on an institutional grant for the REG-NER project at Southeast University, hold memberships in IEEE and ACM, and co-founded A.R.i.S.E. — a nonprofit research institute building collaborative AI research capacity in Bangladesh.
I value reproducibility and open science: most projects ship annotated datasets with DOIs, code repositories, and detailed documentation so others can build on the work.
Research Focus
Sequence Modeling & Financial ML
Transformer architectures and positional encoding techniques such as Time2Vec bring sequence modeling capabilities from NLP into financial time-series forecasting. I apply these methods to the Dhaka Stock Exchange, where market microstructure and data sparsity present distinct challenges. The work includes open dataset releases to support reproducible evaluation.
Low-Resource Bangla NLP
Bangla is spoken by over 230 million people, yet computational resources for its regional variants remain scarce. I work on creating annotated datasets, evaluation benchmarks, and models for dialect machine translation, regional named entity recognition, and hate speech detection — contributing infrastructure that can advance NLP in other low-resource languages as well.
Explainable AI for Imaging
Deep learning models for medical diagnosis and agricultural monitoring are only useful in practice if clinicians and practitioners can understand and trust their outputs. I pair vision transformers and CNN architectures with post-hoc interpretability methods (Grad-CAM, LIME, SHAP) for brain MRI analysis, diabetic retinopathy screening, and plant leaf disease detection.
Education
Feb 2022
- Thesis work on transformer-based deep learning for financial time-series prediction.
- Coursework spanning algorithms, machine learning, computer vision, software engineering, and numerical methods.
Experience
- Supervise undergraduate research teams in NLP, time-series forecasting, and XAI; work routinely reaches publication in IEEE and Springer venues.
- Serve as Co-Investigator on the IRT/SEU institutional research grant for the REG-NER project (regional Bangla NER corpus).
- Teach six core undergraduate courses; integrate practical tooling — Git/GitHub, CI, documentation standards — into coursework.
- Contributed to 20+ publications across IEEE conferences, Springer volumes, and journals as supervisor/co-author.
- Co-founded to create structured research mentorship infrastructure for emerging researchers in Bangladesh lacking institutional support; partially self-funded by founders.
- Leading the Financial ML Studies research group — coordinating time-series and emerging market forecasting research across members.
- Serve as Treasurer: managing grant disbursements, conference fee processing, server and operational costs, and budget reconciliation for the institute.
- Coordinating advisor recruitment from senior faculty; connecting junior researchers with publication pipelines across IEEE, Springer, and Q1 journal venues.
- Taught software development (MVC architecture, C#/.NET) and advised final-year project groups.
Feb 2022
- Developed ERP feature modules from requirements through release; built secure report generation and integrated third-party APIs.
Grants & Funding
Featured Projects
Selected Publications
Open Datasets
All datasets are publicly archived on Mendeley Data with permanent DOIs.
Awards & Honours
Memberships & Affiliations
Language Proficiency
Teaching
Department of CSE · Southeast University · Dhaka (Oct 2022 – present)
Student Resources & Work
I try encouraging students to interact with online specially open source materials. i encourage them to build and work on GitHub while myself trying to improve myself and use the tools.
Numerical Methods Assignment Library ↗ Embedded Systems Lab General Codes ↗