Developed an AI-Native travel app allowing users to find places nearby or for an upcoming, love their favorites, plan itineraries, and more! Available on the iOS app store now. Built the front end with Swift. Hosting the app's web service on AWS cloud services with a PHP backend / MySQL database. Utilizing a number of 3rd party API services such as OpenAI and Foursquare Places API. AI Chat completions functionality and Agentic AI functionality for place autocompletion, content moderation, and more - read more here.
Deep Learning backed real estate price predictor. 10 years of historical home sales for the Indianapolis region are fed into a Neural Network model designed to predict for each home in the region: the current value, the value 2 years from now, and the value 5 years from now. These trends are then displayed on a heat map interface. Front end in Javscript/HTML/CSS. ML model uses Pytorch & Pandas on AWS hosted GPUs. ETL done on AWS services like Glue and Sagemaker. CSV, GeoJSONs hosted on S3 instead of any database.
Associated with chihacknight.org, open-source data platform to make civic data available to the public and accessible by anyone. We use the 'Open States Scraper' that pulls government data from a number of online sources. We put the data in an append-only log / blockchain format for ingestion. My most recent contributions are toward Bluesky bots that automatically send out legislative updates - built mostly in Typescript. I'm a small contributor along with a lot of other great people in the Chi Hack Night community, both tech and community-driven non-tech folks.
Attended Llama 4 Hackathon in Seattle and competed for cash prizes and bragging rights to see who could build the coolest project with their latest AI model in just over 24 hours. Our team didn't win, but it was a blast I think we built a decent product given the constraints. We built an accessibility tool for the visually impaired to generate audible descriptions of any video. Front end in JS, backend in Python / run using Node JS. Integrated with Llama API and integrated Whisper model locally for transcription.
