A few builds I’m proud of — a live ML app, a realtime fraud pipeline, an end-to-end data pipeline, and a privacy-first RAG system in the works.
In prerelease
A security- and privacy-first RAG system for your personal files — answers questions over your own documents without handing them to a third party. Currently in prerelease while I finish the branding.
Download & GitHub coming soonLive app · Machine learning
An audio-classification model that identifies 19 Floridian frog species from sound input — record a call, get a species. Built around a deep-learning model with a live web interface.
Open the live app ↗Data engineering · Streaming
A streaming pipeline that scores transactions for fraud signals in near real time, with downstream monitoring. The latest iteration of a project I’ve rebuilt several times to push latency and precision.
View on GitHub ↗End-to-end pipeline
A complete, simulated e-commerce site: bots role-play as customers, dynamic-pricing agents adjust prices on the fly, and performance analytics track it all — generating live data for an end-to-end pipeline you can explore in the architecture diagram. Retired now, but still worth a look.
View the architecture diagram ↗More on GitHub →