HeuriGuard -A Hybrid Machine Learning System for Malicious URL Detection and SMS Spam Classification
Description:
HeuriGuard is an intelligent cybersecurity system designed to detect malicious URLs, phishing links, and SMS spam messages using a hybrid rule-based and machine learning approach. The system analyzes both textual content and URL structural patterns to identify potential threats and provide real-time safety predictions.
Key Features:
- Detection of malicious URLs and phishing links
- SMS spam classification using machine learning
- Natural Language Processing with TF-IDF vectorization
- URL lexical and structural feature analysis
- Hybrid rule-based + ML detection framework
- Web application and Android mobile application support
Technologies Used:
Machine Learning
- Gradient Boosting
- Naïve Bayes
- Scikit-learn
- NLP
- Tokenization
- TF-IDF vectorization
- Text feature extraction
Development
- Python
- Flask
- HTML
- CSS
- JavaScript
Deployment
- Backend hosted on Render
- Custom domain deployment
Live Demo:
🌐 https://heuriguard.mdjaveedkhan.me
Outcome:
This project demonstrates how machine learning can be integrated with web and mobile technologies to create practical cybersecurity tools capable of identifying malicious digital content.
FastAPI Food Delivery Backend System
Description:
FastAPI Food Delivery Backend System is a real-world backend application designed to manage restaurant menu operations, customer orders, cart workflows, and delivery processing. The system is built using FastAPI and demonstrates how modern backend systems handle API requests, validation, and multi-step workflows efficiently. It simulates real-world food delivery platforms by integrating core backend functionalities such as CRUD operations, filtering, search, sorting, and pagination.
Key Features:
- RESTful API development using FastAPI
- Menu management with full CRUD operations
- Cart system with add, update, and checkout workflow
- Order processing with real-time validation
- Pydantic-based request validation and error handling
- Search functionality (case-insensitive keyword search)
- Sorting (price, category, name)
- Pagination for efficient data handling
- Combined browse endpoint (search + sort + pagination)
- API testing using Swagger UI
Technologies Used:
Backend Development
- Python
- FastAPI
- Pydantic
- Uvicorn
API Features
- CRUD Operations
- Data Validation
- Helper Functions
- Multi-step Workflow (Cart → Checkout → Order)
- Search, Sorting, Pagination
Outcome:
This project demonstrates strong backend development skills by implementing real-world API workflows, data validation, and system design using FastAPI. It helped in understanding how scalable backend systems are structured and how different components interact in a production-like environment.
Malicious URLs Detection Using Machine Learning Techniques
Description:
Developed a live Malicious URL Detection system capable of identifying phishing and malicious URLs in real-time. The system utilizes machine learning algorithms to analyze URL features and classify them as safe or malicious.
Technologies Used: Python, Scikit-learn, Flask, REST API, Render (for deployment)
Key Features:
- Real-time URL analysis and classification
- Integration with a web interface for user interaction
- Deployment on Render for scalability and accessibility
Outcome:
Successfully deployed the Malicious URL Detection system, providing users with a reliable tool to identify and avoid phishing and malicious URLs, enhancing online security.
Text-to-Image Generation Model
Description:
- Developed Generative AI web application converting text prompts into high-quality images using Stable Diffusion model
- Optimized image generation pipeline reducing latency by 40% through efficient preprocessing and caching mechanisms
- Secured 3rd place at IGNITE HACK 2024 among 50+ competing teams, demonstrating innovation and technical excellence
Technologies Used:
Python, TensorFlow, Stable Diffusion, Google colab
Outcome:
Successfully developed a Generative AI model, enabling users to generate high-quality images from text prompts, showcasing the power of AI in creative applications.