About
About
I'm Ray, a computer scientist based in North Carolina. I build projects of all sizes, and when I want to probe an idea, I build small: just enough to find out whether a new model, tool or piece of hardware can actually do the thing I hope it can.
AI coding agents are a big part of why. An MVP that would have taken me months in 2020 now takes days, so the cheapest way to test an idea is often to build it. Most of those builds don't go anywhere, but each one teaches me something. This blog is where I write those lessons up.
What I'm working on now
- A history guessing game run by an on-device AI model. I'm testing what Gemini Nano on a Pixel can and can't do with Wikipedia articles, and designing the game around those limits.
- Pod, a Discord bot that groups members of a server into small matched trios, then measures whether they choose to keep talking.
- Election data side quests, like checking whether a North Carolina congressional race would have flipped under the previous district map.
Background
I have a BS in Computer Science (cum laude) and an MS in Computer Science with a concentration in Data Science and Visualization, both from Appalachian State University.
Most of my work has sat where data meets the web:
- Climate and weather tools. As a research assistant in Appalachian State's Department of Geography and Planning, I built a full-stack web app for running the FLEXPART air trajectory model and an ETL pipeline that collects hourly NOAA weather station data. I later built a Vue and Node app for exploring and downloading data from mountain weather stations.
- Political data. I interned on the data and tech team at Movement Labs, working on a campaign reporting app and testing a voter data pipeline. I also built a Tableau dashboard of North Carolina voter turnout by demographic group from my own voter file pipeline.
- Front-end contract work in Vue.js for Logicom USA.
- Personal builds, including a video aggregation platform (Nuxt, Express, PostgreSQL, Redis and BullMQ) and an MLB analytics pipeline built on pitch, venue and weather data.
Tools I reach for
Python, TypeScript and JavaScript; Vue/Nuxt, React and Node; PostgreSQL, SQLite and Redis; pandas and scikit-learn; Docker. Lately, coding agents and small local models such as Gemma and Gemini Nano.
Get in touch
- GitHub: See Projects
- Email: Professional contact

