Designing and developing smart, accessible software solutions while conducting research on innovative technologies and their real-world applications.
Oxford, Ohio, USA
Dhaka, Bangladesh · On-site · Full-time
Software Engineer: Nov 2023 - Jul 2024
Associate Software Engineer: Oct 2022 - Oct 2023
Tech Stack: Java Spring Framework, Hibernate, Oracle SQL, JSP, JavaScript, jQuery
Dhaka, Bangladesh · Full-time · On-site
Tech Stack: Android Application Development, RTOS
Nov 2024 – Present
Worked on Java Spring Boot API to generate wheelchair-accessible sortest routes based on latitude-longitude queries. Returns routes with surface type, incline, turns, and duration for optimized accessibility.
Visit our pageJan 2023 – Apr 2023
A full-stack, role-based online store application built with Java, Spring, and Hibernate, which uses Docker for containerized deployment and a PostgreSQL database to manage e-commerce workflows for various user roles.
Project link2022
Developed a contacts app for Rajshahi University using Android Studio and Kotlin. Implemented MVVM architecture for better maintainability and scalability.
Google play linkNov 2024 – Dec 2024
Processed GeoTIFF satellite images and OpenStreetMap (OSM) data for analysis. Applied zero-shot learning with GPT-4o to detect crosswalks from high-resolution GIS imagery without labeled training data, achieving 97.5% accuracy.
Paper linkJan 2023 – Apr 2023
Implemented order processing, table management, and user profile management with secure login. Ensured security with input validation, custom exception handling, and request filtering.
Project linkMarch 2025
Developed a deep learning model to recover thin skeletons from thick, noisy road network images from OpenStreetMap (OSM). Used U-Net with residual blocks, iterative thinning, and multi-task learning for accurate skeletonization.
Project linkOct 2024 – Nov 2024
This project detects human activities using form IMU sensor data. The dataset includes 15 users performing various activities. Features are extracted using a sliding window approach, and multiple models (Random Forest, XGBoost, CNN, Logistic Regression) are trained and evaluated. Random Forest, XGBoost, and CNN performed the best.
Project linkNov 2024 – Dec 2024
Analyzed urban sprawl in Dhaka using remote sensing and GIS with USGS data. Computed and visualized NDBI, SAVI, MNDWI, and IBI indices for land use assessment and spatiotemporal analysis of urban growth.
Project linkOxford, Ohio, USA
Rajshahi, Bangladesh
Rajshahi, Bangladesh
Authors: S. Karki, E. Han, N. Mahmud, S. Bhunia, J. Femiani, and V. Raychoudhury
Type: Workshop
Authors: ASM Mobarak Hossain *, Nadim Mahmud *, Ethan Han, Neil Advani, Bibodh Baral, Md Osman Gani, Vaskar Raychoudhury
Type: Workshop
McVey Data Science Building 364, Miami University
Oxford, OH 45056
United States
Monday – Friday
9:00 AM – 5:00 PM PST