Tucson, AZ · keatonwilson@me.com · LinkedIn · Twitter · Scholar
I build data-driven products that connect analysis to real decisions — from Shiny dashboards and interactive visualizations to LLM-powered pipelines and agentic workflows. Background in ecology and environmental science; now applying the same instinct for finding signal in noise to software and AI.
- Solutions Development — End-to-end data products, dashboards, and tooling at the intersection of research and business
- Data Science — Statistical modeling, machine learning, and deep expertise in R and Python
- AI Engineering — Building with the Anthropic/Claude API, LangChain orchestration, RAG pipelines, and agentic workflows
cartoTaco · SvelteKit Supabase Mapbox Tailwind
Full-stack interactive map app for exploring Tucson’s taco scene — marker clustering, multi-stop route planning, k-NN taste recommendations, spot comparison mode, and user auth.
cartoTacoMenuExtract · Python Claude Vision Streamlit Supabase
AI-powered admin pipeline that extracts structured menu data from restaurant photos using Claude Vision, with a staging review UI and direct promotion to the production database.
Data & AI
Started coding in grad school to model how insects decode plant volatile signals — essentially reverse-engineering nature's data pipelines. That same drive to find signal in noise and build systems around it is what I bring to data science and AI engineering today.
Currently at KS&R