A collection of projects showcasing my work in web development, machine learning, and more.
Multi-agent AI system that detects customer issues across internal channels and social media, routes incidents to teams, and generates reverse prompts guiding employees on next steps. Read-only pipeline with RAG and guardrails.
Brain-computer interface system for classifying imagined Arabic speech from EEG data. Two-stage pipeline using transformer neural networks for word classification and LLM for sentence reconstruction.
SEO-optimized public website for GC International, built with Astro, React, Airtable CMS, and Three.js for interactive 3D animations. Content managed via Airtable and deployed to Vercel.
Multi-stage phishing detection browser extension with real-time URL analysis and ML classification. Features USI algorithm with FAISS similarity search and Logistic Regression classifier achieving 99.66% accuracy.
A full-stack anime recommender system representing a multi-stage hybrid recommendation engine that adapts to each user by combining non-personalized popularity, content-based filtering, and collaborative filtering.
CloudDesk is a serverless IT helpdesk built on AWS with role-based employee and agent workflows. It supports secure sign-in, fast ticket routing, and end-to-end lifecycle tracking with real-time status visibility.
Educational demonstrations of social engineering attacks including GUI confusion phishing and privacy dark patterns. Features fullscreen fake login pages and predatory privacy settings UI studies.
Full-stack vehicle auction aggregator processing 50K+ daily lots from Copart, IAAI, Emirates Auction and more. Features VIN search, bid history tracking, advanced filtering, and secure payment gateway integration.
Dynamic BMI-driven meal planning platform with personalized recommendations. Features caloric needs calculation, meal categorization (high-protein, balanced, plant-based), and responsive UI.
Deep learning-based malware detection system using Convolutional Neural Networks. Implements ResNet18 and custom 3-layer CNN architectures with data augmentation for classifying malware samples from binary images.
ML-based fault detection for grid-connected photovoltaic systems. Analyzes electrical and environmental parameters using Decision Trees, k-NN, SVM, and Neural Networks achieving 99.84% accuracy.
Machine learning model for predicting heart diseases in high-risk patients. Uses patient health metrics and clinical data to identify individuals at elevated cardiovascular risk.