Projects Category: Mobile Applications

Project Overview

Eco Up was developed as a graduation project for the Department of Computer Science at Northern Border University. The primary goal was to tackle the accumulation of recyclable waste on campus by shifting student behavior from passive awareness to active participation .

Key Features

• 📱 Smart Scanner: Instant waste recognition using the camera to verify recyclable items.

• 🏆 Gamification Engine: An algorithm that calculates and awards points based on waste type and quantity.

• 🎁 Rewards Marketplace: A digital store where users can redeem points for coupons and discounts.

• 📊 Impact Dashboard: Visual statistics showing personal contribution and environmental impact.

Project Outcomes

• 75% Increase in recycling participation rates during testing.

• 40% Reduction in waste sent to landfills.

• Successfully deployed a scalable architecture capable of handling high user loads with no memory leaks.

GitHub Repo | pub.dev Package

Flutter AI Kit is a feature-rich, all-in-one Flutter library I built to simplify the integration of advanced AI models into mobile applications. Whether you’re working on a chatbot, AI-generated images, or text analysis tools, this package provides everything you need with minimal setup and a clean, modern UI out of the box.

About the Project

With the rising popularity of AI-powered apps, developers often face complex integrations, inconsistent APIs, and UI customization challenges. Flutter AI Kit solves these problems by offering:

• Unified integration with OpenAI, Google Gemini, and HuggingFace

• Beautiful, pre-built chat and image generation UI

• Easy customization and theming to match any app design

This package is ideal for building AI chatbots, virtual assistants, content creation tools, educational apps, or any application that benefits from AI-powered functionality.

Key Features

AI Chat Interfaces

• Pre-configured chat UI widgets for OpenAI and Google Gemini

• Supports conversational state management, user/AI message formatting

• Clean and responsive UI design

AI Image Generation

• DALL·E Integration for AI-generated images

• Includes loading states, error handling, and image previews

Text Processing with HuggingFace

• Text classification, summarization, translation, and more

• Easy integration of any HuggingFace-supported model with REST API support

Modern Flutter Design

• Fully customizable components using Flutter’s built-in theming

• Dark/light mode support

• Rounded shapes, shadows, and responsive design elements

Developer-Friendly

• Simple API key setup for OpenAI, Gemini, and HuggingFace

• Full null safety and type safety

• Modular architecture for easy feature scaling

🔗 Project Links

• 💻 GitHub: github.com/Hamed233/flutter_ai_kit

• 📦 Pub.dev: pub.dev/packages/flutter_ai_kit

In an increasingly connected digital world, cyber threats continue to evolve at an alarming rate. With phishing attacks and malicious links becoming more sophisticated, there is an urgent need for intelligent, real-time protection systems. That’s where this AI and Machine Learning-powered Malicious URL Detection System comes into play—a cutting-edge solution designed to help users stay protected from online threats before they strike.

This project is a robust mobile and backend system developed using modern technologies like Flask (Python), TensorFlow, Firebase, and Flutter, with a core focus on real-time threat detection, speed, security, and usability.

Project Goals

The primary objective of this project is to safeguard users from phishing, malware, and other cyber threats by analyzing URLs in real-time using artificial intelligence and machine learning models. It empowers users with instant threat detection, historical scan tracking, and cybersecurity education.

Key Features & Functionalities

URL Safety Analysis

• Real-Time Detection: Analyze links using a trained LSTM model to detect phishing, malware, and suspicious patterns.

• AI-Based Classification: Uses a deep learning model trained on malicious and safe URL datasets for high accuracy and fast inference.

Scan Results & Insights

• Detailed Scan Feedback: Displays if the URL is safe or malicious, with context on detected risks.

• Scan History: Users can view their previously scanned links with timestamps, helping them stay aware of recurring threats.

Instant Notification Alerts

• Users are immediately notified if a link is classified as dangerous, preventing potential harm before any interaction.

Mobile Application Design

Built with a modern Android architecture, the mobile app is user-centric and secure:

Security-Focused Login

• OAuth 2.0 & JWT Authentication

• Firebase for credential management

• Encrypted session storage

Intuitive User Interface

• URL submission with regex validation

• History view powered by SQLite (Room)

• URL caching with LRU policy for speed

• Dark/light themes and accessible UI

Machine Learning & Model Deployment

LSTM-Based Detection Engine

• Bidirectional LSTM Model with ONNX runtime for performance

• Achieves 98.4% F1-score using datasets from PhishTank and VirusTotal

• Handles requests in 320ms on CPU, 85ms on GPU

Preprocessing & Tokenization

• Custom tokenizer supports Unicode-safe encoding

• Sequences converted for LSTM input with max length of 200

Model Serving

• Deployed using TensorFlow Serving over gRPC

• Supports async requests via Flask API

Backend Architecture

A multi-tier microservices architecture ensures scalability and high performance:

Flask API Gateway

• Blueprint routing and rate limiting (100 req/IP/hr)

• Input sanitization with Bleach to prevent XSS/SQLi

• Secure API request handling using HMAC verification

API Endpoints

• POST /api/v1/scan: Accepts a URL and returns classification

• GET /api/v1/history: Fetches user’s scan history

Security Features

• HTTPS encryption for all client-server communication

• JWT-based session management

Technical Stack

• Frontend (Mobile): Flutter

• Backend API: Flask (Python)

• Authentication: Firebase, OAuth 2.0, JWT

• Machine Learning: TensorFlow, LSTM, ONNX

• Storage: Firebase, SQLite

• Communication: REST + gRPC

Project Overview

Engfull is a cross-platform educational platform that merges gamification with English language learning, offering an engaging and personalized experience for learners of all levels. The project is designed to make language acquisition fun, adaptive, and accessible through mobile and web platforms.

Built with performance, scalability, and security in mind, Engfull supports real-time multiplayer games, dynamic content delivery, and subscription-based access — all managed from a powerful backend dashboard.

Key Features

• Gamified Learning Experience: Boost your English proficiency through exciting mini-games that target vocabulary, grammar, spelling, and comprehension.

• Personalized Challenges: Adaptive exercises and quizzes that adjust to the user’s level, designed to encourage progression and retention.

• Mobile App (iOS & Android): Fully featured app built with Flutter, supporting in-app purchases, real-time games, and smooth offline support.

• Dynamic Web Application: Laravel-powered backend with custom admin and content creator dashboards, multi-auth, and full user management.

• Secure Payment Integration: Seamless subscription management via Tap Payment Gateway for both mobile and web users.

• Robust Security Measures: Includes WAF protection, Google reCAPTCHA, IP rate-limiting, tokenized APIs, and DDoS prevention.

• Real-Time Features & Analytics: Firebase integration for multiplayer games, real-time database, FCM push notifications, and analytics tracking to monitor performance and crash reports.

• Multi-Role Authentication: Separate login and access control for end users and content creators, each with dedicated dashboards.

• Admin Dashboard: A comprehensive control panel to manage users, content, categories, games, subscriptions, and more.

• Creator Dashboard: A dedicated dashboard for content creators to upload learning materials, create exercises, track engagement, and manage student interactions — empowering educators and contributors.
• Group Chats for Each Creator: Interactive group chat rooms where learners can engage with their chosen creators and fellow students, ask questions, receive updates, and build a learning community in real-time.

Technologies Used

• Backend: PHP, Laravel.
• Database: MySQL, Firebase Realtime Database.
• Cloud: Firebase.
• Apps: Flutter (IOS, Android).
• Security: Google reCAPTCHA, WAF, IP Tracking, Token Auth, Rate Limiting
• Others: FCM Notifications, Firebase Analytics, RESTful API, Multi-auth.