Archive // All Projects
A comprehensive index of technical implementations, research experiments, and engineering solutions.
Developed a high-stakes aerospace tracking system using .NET Core and React. I integrated JWT-based authentication for secure access and implemented automated CI/CD pipelines via Azure DevOps. The platform leverages Azure ServiceBus for real-time messaging, ensuring immutable audit trails and supply chain transparency across large-scale international enterprise operations.
Developed 15+ advanced Python scraping scripts to extract millions of business registry records across international markets. I personally managed client communication and engineered solutions to bypass complex anti-scraping defenses for dynamic websites. The system processed up to five million data points in multiple languages, ensuring high integrity through custom post-processing automation.
Engineered a distributed web scraper using Python and MPI to harvest GitHub Trending data across multiple nodes. The system utilizes CloudScraper to bypass advanced anti-scraping measures while maintaining 99.9% data integrity through automated deduplication. Developed as part of my Western University coursework, it significantly optimizes large-scale data extraction throughput.
Developed an AI-powered agricultural diagnostic tool using Convolutional Neural Networks to detect 10 distinct tomato leaf diseases. The system achieves 96.3% accuracy using a custom architecture optimized for resource-constrained devices. It also benchmarks performance against a fine-tuned ResNet50 model, offering a scalable solution for real-time crop health monitoring and sustainable farming.
Neo-Atom is a sophisticated, privacy-first AI assistant engineered to bridge the gap between local LLMs and Windows OS automation. Leveraging an agentic tool-calling framework, it processes natural language through high-speed local STT and TTS engines to execute complex system commands, web searches, and safe mathematical computations autonomously.
Engineered an advanced taxi fare prediction system using the New York City dataset as part of my Master’s AI coursework. I benchmarked various models, selecting LightGBM for its superior performance, achieving an 83.9% R² score. The project utilized Docker to automate travel distance and time extraction from complex map data.
This Master's directed study research evaluates the ethical alignment of prominent LLMs, including ChatGPT, Gemini, and Claude, across fifteen curated dilemmas. I transformed qualitative AI responses into a quantitative binary matrix for statistical analysis. Utilizing Cochran’s Q and Fleiss’ Kappa tests, I established a rigorous framework to measure inter-model agreement and behavioral consistency.
Engineered a high-end, full-stack photography portfolio using Next.js 16, React 19, and Tailwind CSS v4, paired with a robust .NET 8 backend. The platform features an immersive visual experience with smooth Lenis scrolling, complex Framer Motion animations, and a secure, bespoke dynamic image management system with gallery integrated with Adobe Lightroom.
Architected a high-performance, polymorphic portfolio platform from scratch using Next.js 14 and TypeScript. The system features a custom state-driven "Mode Switcher" that recontextualizes the entire UI—from hero text to project filtering—based on a viewer's interest. I utilized Gemini Stitch for UI design inspiration, ensuring a modern, developer-centric aesthetic with 99.9% performance scores.
E-commerce Platform
Developed a high-performance e-commerce platform using .NET 7 and React during my time at FSP Media. I integrated Stripe for secure payments and utilized Redis for efficient caching to handle high traffic. The entire application was containerized with Docker, ensuring scalable and consistent deployments across production and development environments.
Developed a comprehensive cross-platform e-commerce application using Flutter, Node.js & MongoDB. The system replicates core Amazon functionalities, including secure user authentication and integrated mobile payments via Google Pay. It also features a dedicated seller dashboard for tracking inventory and real-time sales statistics, demonstrating a complete full-stack mobile commerce solution.
