Early Dengue Detection in Bangladesh Using Symptom Analysis and Time-Series Forecasting (Research Based Project)
This research based project aims to detect dengue early in Bangladesh using two datasets: district-wise dengue cases (2016–2024) and a symptom-based survey. A two-step machine learning model combines time-series forecasting (using Prophet) and symptom-based classification (using Random Forest and XGBoost). The approach achieved up to 92.5% accuracy, offering a non-invasive and efficient method for early dengue diagnosis.
Used Python, LSTM, Prophet, Scikit-learn, XGBoost, Random Forest, Google Forms for data collection
Privacy-Preserving Early Detection of Autism Spectrum Disorder Using Federated Learning (Research Based Project)
This project introduces a federated learning-based approach for early detection of Autism Spectrum Disorder (ASD) using the Autism Gait and Gesture dataset. By implementing a client-server architecture, the model ensures decentralized training while preserving patient privacy. The proposed system achieved 97.5% accuracy, demonstrating its potential for secure, accessible, and accurate ASD diagnosis.
Used Python, Deep Learning Models, RSA Encryption, Tensorflow
Paper and Pencil Games (Android Game)
This project features a collection of classic paper and pencil games like Tic Tac Toe, Dots and Boxes, SOS, SIM, Pong Hue Ki, and Four in a Row for simple, offline fun. Designed for all ages, the app promotes strategic thinking, cognitive development, and social interaction through intuitive gameplay. It's free to use, easy to access, and ideal for both quick breaks and longer play sessions.
Used Unity, C#, Firebase and Admob
Guti – Traditional Strategy Game (Android Game)
This project brings the classic South Asian board game “Guti” to life through an engaging and intuitive mobile app. It features multiple traditional modes including 3 Guti, 6 Guti, 16 Guti, Lau Kata Kati, and Pretwa, supporting both single-player (AI) and multiplayer gameplay. Designed with adjustable difficulty, stunning visuals, and smooth user interaction, the game appeals to players of all ages and skill levels. A leaderboard and achievement system encourages competition and replayability. The app preserves cultural heritage while offering a modern, accessible experience through digital innovation.
Used Unity, C#, Graphics Designing Tools, Firebase and Admob
Computer Graphics Projects (Kitchen, Resort Backyard)
This academic project uses C/C++ along with OpenGL, GLU, and GLUT libraries to explore 3D computer graphics. It applies key techniques like translation, rotation, shading, lighting, and texturing to create and manipulate realistic models, including a resort backyard. The project showcases a strong foundation in graphical rendering, simulation, and visualization.
Used OpenGL libary and C/C++
Puzka- Personal Assistant App (Android App)
PUZKA is a personal assistant Android app designed to help users manage daily tasks. It features a notepad, transaction manager, routine organizer, drawing tool, and an image-to-text recognizer. The app also offers biometric authentication for security and uses Firebase ML Kit for smart text recognition.
Used Java, OpenCV, Firebase ML Kit, SQLite, Biometric Authenticator, Android Studio
eTutor (Android App)
eTutor is an Android app that connects students with qualified tutors for various subjects, offering an easy-to-use platform for browsing, communication, and booking sessions. The app leverages Google Maps API for location-based searches and ensures trust through profile verification to prevent fraud. Built with Java, SQLite, Firebase, and Google Maps API, it provides an efficient and secure way for students and tutors to connect.
Fitness Freak Foodie (Android App)
Fitness Freak Foodie is an Android app designed to help users track their physical fitness journey by recording data such as calories burned, BMI, and BMR. The app offers a user-friendly interface and integrates Firebase for data storage and synchronization. Developed using Java and Android Studio, it provides an easy way to monitor and manage fitness goals.
Tank Battle Game (Windows Game)
Tank Battle Game is a two-player real-time action game developed using Unity and C#. It supports both online and offline modes, with AI opponents powered by Minimax, Heuristic Search, and Reinforcement Learning algorithms. The project focuses on strategic gameplay, offering users an engaging and intelligent battle experience.