Word-Maestro

Game

An engaging educational word game built with C++ and iGraphics library that challenges players' vocabulary and spelling skills. Traditional vocabulary learning methods can be monotonous and fail to maintain learner engagement. Word Maestro addresses this by turning word learning into an interactive and entertaining experience.

Problem Statement:

Traditional vocabulary learning methods can be monotonous and fail to maintain learner engagement. Word Maestro addresses this by turning word learning into an interactive and entertaining experience.

Key Objectives:

• Create an engaging platform for vocabulary enhancement
• Provide immediate feedback on word knowledge
• Implement multiple difficulty levels to suit different skill levels
• Track player progress through a high-score system
• Deliver an intuitive and visually appealing user interface

Key Features:

• Multiple game modes
• Difficulty selection
• High score tracking
• Sound effects and background music
• Visual character representation
• Pause/Resume functionality
• Interactive keyboard input
• Performance tracking

Results & Impact:

• Successfully developed and deployed a functional word game.
• Created an engaging learning tool for vocabulary enhancement.
• User engagement through high-score system and positive feedback.
• Provides educational value through interactive learning and entertainment through gamification.

Key Learnings:

• Graphics programming with OpenGL
• Game development principles
• User interface design
• File system management
• Audio integration in applications

Tech Stack: C++, iGraphics, irrKlang, STL, OpenGL, Windows API

Resort Administration System

Desktop Application

A comprehensive Java-based resort management system for booking accommodations, tours, venues, and managing customer services. Modern resorts face complex challenges in managing multiple services, accommodations, and customer interactions efficiently. Manual booking systems and fragmented management tools lead to operational inefficiencies, booking conflicts, and reduced customer satisfaction.

Problem Statement:

Modern resorts face complex challenges in managing multiple services, accommodations, and customer interactions efficiently. Manual booking systems and fragmented management tools lead to operational inefficiencies, booking conflicts, and reduced customer satisfaction.

Key Objectives:

• Streamline the booking process for cottages, venues, and tours
• Provide a unified administrative interface for resort management
• Enhance customer experience through easy registration and service access
• Implement secure authentication for both customers and administrators
• Manage diverse resort amenities including gym passes and health support services

Key Features:

• User Management (Customer & Admin)
• Booking Systems (Cottages, Venues, Tours, Transport)
• Financial Operations (Payment, Withdrawal, Checkout)
• Amenity Management (Gym, Health Support)
• Multi-modal Transportation Options
• Location Variety (Beach, Forest, Mountain, Island, Town)
• Administrative Tools (Dashboard, Tracking, Monitoring)
• Media Integration (Audio, Slides, Images)

Results & Impact:

• Fully functional resort management system with integrated booking for multiple services.
• Streamlined administrative processes and enhanced customer service capabilities.
• Reduced administrative overhead, improved booking efficiency, and better resource utilization.

Key Learnings:

• Complex system integration techniques for diverse resort services.
• User interface design principles for administrative and customer-facing applications.
• Database management practices for handling bookings, user data, and financial transactions.
• Security implementation methods for protecting sensitive information.

Tech Stack: Java, Java Swing, JDBC, MySQL, JaCo MP3 Player, Custom UI Components (.form files), JPA/Hibernate

DecorateMyNest (Java)

Desktop Application

A comprehensive interior design project management system built with Java and SQL Server. It is designed to revolutionize how interior design businesses manage their projects, resources, and client relationships.

Problem Statement:

Interior design businesses often struggle with managing projects, resources, and client relationships efficiently using manual or disparate tools, leading to workflow inefficiencies and communication gaps.

Key Objectives:

• Streamline project workflow from initial consultation to completion.
• Centralize management of projects, clients, resources, and finances.
• Ensure data integrity and reliability through robust database design.
• Provide a user-friendly interface requiring minimal training.

Key Features:

• Comprehensive Project Management (cataloging, cost estimation, progress tracking, archiving, review system)
• Client Management (detailed profiles, appointment scheduling, project reservations)
• Resource Management (material inventory tracking, vendor relationships, employee roster/scheduling)
• Financial Management (transaction tracking, payment processing, financial reporting)

Results & Impact:

• Delivered an efficient and organized system for interior design project management.
• Enabled streamlined workflows, centralized data, and scalable resource handling.
• Provided a reliable and user-friendly tool for interior design businesses.

Key Learnings:

• Application of three-tier architecture in desktop application development.
• GUI design using Java Swing for intuitive user interaction.
• Database design and management with SQL Server and JDBC.
• Importance of modular design for maintainability and scalability.

Tech Stack: Java, Java Swing, SQL Server, JDBC, NetBeans

ReUse

Mobile App (Android)

A sustainable Android application for donating and reusing unwanted items within communities. It addresses the issue of usable items ending up in landfills by connecting donors with those in need.

Problem Statement:

Many usable items are discarded, contributing to waste, while individuals and communities could benefit from them. Existing platforms lack efficiency in connecting local donors and recipients for item reuse.

Key Objectives:

• Facilitate easy donation and discovery of unwanted items within local communities.
• Reduce waste by promoting the reuse of items.
• Build a community-driven ecosystem for sustainable resource sharing.
• Implement secure user authentication, real-time messaging, and location-based discovery.

Key Features:

• User Authentication (Email, Google Sign-in, Profile Management)
• Item Listing (multi-image upload, category selection, location tagging, description)
• Search and Discovery (category filters, text search, location-based item discovery using Google Maps)
• Real-time Messaging System (chat history, notifications)
• Material Design UI with custom fonts and responsive layouts.

Results & Impact:

• Successfully developed a functional Android platform for item donation and reuse.
• Implemented core features including authentication, item listing, chat, and location services.
• Contributes to environmental sustainability by reducing waste and promotes community resource sharing.

Key Learnings:

• Best practices in Android app development using Java and XML.
• Integration and management of Firebase services (Auth, Realtime DB, Storage).
• Implementation of Google Maps API for location-based features.
• Real-time database management and synchronization for features like chat.
• Feature prioritization and user experience design for mobile applications.

Tech Stack: Java, XML, Firebase (Auth, Realtime DB, Storage), Google Maps SDK, Google Places API, Picasso, Dexter, AutoImageSlider

ScholarSphere

Web Platform

A comprehensive university admission and academic community platform for Bangladesh, addressing the fragmented and complex admission process.

Problem Statement:

The university admission process in Bangladesh is complex and fragmented, making it difficult for students to track timelines, find accurate information, and connect with the academic community for guidance.

Key Objectives:

• Create a centralized platform for tracking university admissions and information.
• Provide verified and accurate details about universities and programs.
• Foster an academic community for knowledge sharing and support.
• Implement a trusted profile validation system for authenticity.

Key Features:

• University Profiles System (comprehensive info, faculty, requirements, maps)
• Live Admission Tracker (real-time status, deadlines, results, quotas)
• User Management System (Role-based access: Admin, Student, Alumni; Profile Verification; Secure Auth)
• Academic Discussion Forum (community posts, knowledge sharing, moderated content)
• Trusted Profile Validation (Admin-managed, document-based identity verification)

Results & Impact:

• Developed a centralized platform for university admission information and a verified academic community.
• Streamlined admission tracking and facilitated trusted information sharing.
• Provided easy access to admission info for students and improved visibility for universities.

Key Learnings:

• Importance of user verification and data authenticity in educational platforms.
• Value of structured information presentation for complex data like university admissions.
• Community-driven content moderation strategies.
• Significance of real-time updates in time-sensitive processes.
• Web development using PHP, MySQL, and Bootstrap for a responsive platform.

Tech Stack: PHP, HTML, CSS, JavaScript, Bootstrap, jQuery, Font Awesome, SendGrid, MySQL, Composer

DecorateMyNest (.NET React)

Full-Stack Web App

A full-stack web application for managing interior design projects, appointments, and client interactions, built with .NET and React. It addresses challenges in managing multiple projects, client communication, scheduling, and financial tracking.

Problem Statement:

Interior design firms struggle with efficiently managing multiple projects, client interactions, appointments, and financial transactions, often leading to communication gaps, scheduling conflicts, and tracking difficulties.

Key Objectives:

• Streamline the interior design project management workflow.
• Facilitate seamless communication between clients and designers.
• Automate appointment scheduling, project tracking, and financial management.
• Provide real-time project status updates and manage inventory/vendors.

Key Features:

• User Management (Auth, Role-based access: Admin, Client, Employee, Profiles)
• Project Management (Creation, Tracking, Status updates, Task assignment, Archival)
• Appointment System (Client consultation scheduling, Room allocation, Status Tracking, Employee availability)
• Financial Management (Cost estimation, Installment payment tracking, Transaction history, Reporting)
• Client Portal (Booking, Status view, Payments, Communication) & Admin Dashboard (Overview, Resource allocation, Catalog management)

Results & Impact:

• Successfully implemented a full-stack solution improving project management efficiency for interior design firms.
• Enhanced client satisfaction through better communication and real-time updates.
• Streamlined financial tracking, appointment scheduling, and resource management.

Key Learnings:

• Full-stack development with .NET Web API (C#) and React (JavaScript/JSX).
• Importance of proper state management (e.g., project states, payment states) in complex applications.
• Value of clear separation of concerns in a three-tier backend and component-based frontend architecture.
• Benefits of using DTOs, repository pattern, and factory pattern in backend development.
• Frontend development with custom hooks, centralized API service, and responsive design using Tailwind CSS.

Tech Stack: C# (.NET), ASP.NET Web API, Entity Framework, AutoMapper, JWT, React, Vite, React Router, React Bootstrap, Tailwind CSS, SCSS, Axios, React Hook Form, Yup

Advanced ML Research Portfolio

AI/ML Research

A collection of ML research projects addressing challenges in: 1. Bangla Sentiment Analysis (comparing LSTM, GRU, RNN). 2. Diabetic Retinopathy Detection (multi-feature extraction, multi-class classification). 3. Weather Pattern Analysis (neural network regression for temperature prediction). 4. Sign Language Classification (Bangla sign language recognition).

Problem Statement:

This portfolio tackles diverse ML challenges: low-resource NLP for Bangla, accurate medical image classification for Diabetic Retinopathy, complex regression for weather prediction, and accessibility through Bangla sign language recognition.

Key Objectives:

• Develop and compare neural architectures (LSTM, GRU, RNN) for Bangla sentiment analysis.
• Implement multi-feature extraction and ensemble methods for Diabetic Retinopathy detection.
• Create accurate weather prediction models using neural network regression.
• Advance Bangla sign language recognition technology for improved accessibility.

Key Features:

• Bangla Sentiment Analysis: Text preprocessing, custom tokenization (BNLP), word embeddings, multi-architecture model comparison.
• Diabetic Retinopathy Detection: Multi-feature extraction (gray, color, Sobel, SIFT), image preprocessing, ensemble classification (SVM, RF, KNN).
• Weather Pattern Analysis: Data normalization, neural network regression for temperature prediction, multi-variable analysis.
• Sign Language Classification: Real-time processing considerations for Bangla sign language.

Results & Impact:

• Contributed insights into model performance for low-resource NLP (Bangla).
• Developed effective feature extraction and classification pipelines for medical imaging.
• Demonstrated neural network capabilities for complex weather data regression.
• Advanced research in accessibility technology for Bangla sign language.

Key Learnings:

• Comparative analysis of different neural network architectures for NLP tasks.
• Techniques for multi-feature extraction and ensemble learning in computer vision.
• Application of deep neural networks for regression problems with complex, non-linear data.
• Challenges and solutions in processing low-resource languages and developing accessibility-focused ML models.
• Data preprocessing, normalization, and feature engineering for diverse datasets.

Tech Stack: Python, PyTorch, Keras, NumPy, Pandas, OpenCV, scikit-image, scikit-learn, BNLP, Google Colab, Jupyter Notebooks

Smart Video Doorbell

IoT/AI

An IoT-based smart doorbell system with face recognition, mobile app control, and real-time notifications. It addresses the lack of modern security features and remote monitoring in traditional doorbells.

Problem Statement:

Traditional doorbells lack modern security features like video surveillance, facial recognition, and remote access control, leaving homeowners unable to verify visitors or manage access remotely.

Key Objectives:

• Create a secure and reliable smart doorbell system with real-time video streaming and face recognition.
• Provide convenient mobile app control for monitoring and access management.
• Enable secure access through multiple authentication methods (PIN, Face, Remote App).
• Deliver instant notifications for doorbell events, motion detection, and recognized/unknown faces.

Key Features:

• Real-time Video Streaming (MJPEG to Android app)
• AI-powered Face Recognition (on Raspberry Pi using face_recognition library)
• Multiple Authentication Methods (PIN code via keypad, Face Recognition, Remote App Unlock)
• Mobile App Control (Kotlin, Jetpack Compose for monitoring, access, user management)
• Real-time Notifications (Firebase Cloud Messaging for doorbell rings, motion, face alerts)
• Hardware Control (ESP32 for doorbell, lock, keypad; MQTT for communication)
• User face enrollment and training mode.

Results & Impact:

• Developed a functional smart doorbell system with high face recognition accuracy and low video streaming latency.
• Provided an intuitive mobile app for remote monitoring and secure access control.
• Enhanced home security and convenience for users through real-time alerts and multiple authentication options.

Key Learnings:

• Coordinating distributed IoT devices (Mobile App, Raspberry Pi, ESP32) using MQTT and Firebase.
• Optimizing real-time video processing (MJPEG streaming) and face recognition on resource-constrained devices.
• Implementing secure access control with multiple authentication factors in an IoT environment.
• Mobile app development with Kotlin and Jetpack Compose, including state management for real-time updates.
• Hardware-software integration challenges and security-first development approach in IoT systems.

Tech Stack: Kotlin (Android Jetpack Compose), Python (face_recognition, OpenCV, Firebase Admin SDK, Paho MQTT, PiCamera2), C++/Arduino (ESP32), Firebase (Messaging, Admin SDK), Retrofit, MJPEG, MQTT, Raspberry Pi, ESP32