Denials Tracker

I designed, developed, and implemented a progressive web application to streamline the investigation and resolution of insurance claim denials. Deployed locally on-premise at our clinic, the application enabled the recovery of over $60,000 in previously denied claims within its first year of use. The application utilizes the SvelteKit frontend framework, Skeleton UI component library, Supabase for backend services, and AG Grid Community for advanced data grid display. Both the frontend and backend are configurable for online or on-premise hosting, ensuring that protected patient data remains securely within the clinic environment.

AI Discord Bot

This ongoing project serves as my experimental platform for integrating the latest advancements in open-source and self-hosted artificial intelligence. Developed in Python using discord.py and the Discord API, the Discord bot implements many AI functionalities, including a large language model (LLM) powered chatbot, Stable Diffusion image generation, AI-powered text-to-speech, and many more experimental features. The project is continuously updated as I explore and implement new technologies, making it a dynamic showcase of my skills in AI integration and Discord bot development.

SAM2 Image Segmentation & Replacement

This proof of concept project is a Python web app that lets users interactively segment and replace regions in images. Built with Gradio and powered by Meta's Segment Anything Model 2 (SAM2), users can upload images, draw a rough mask, and have the app refine it for precise segmentation and seamless region replacement. The project demonstrates skills in Python, deep learning model integration, computer vision (OpenCV), and rapid UI development with Gradio.

Portfolio Website

This website showcases my portfolio, built with Svelte and TypeScript. It features a responsive design and a clean layout with Tailwind CSS to highlight my projects and skills. The site is optimized for performance and accessibility, ensuring a great user experience across devices.