I’m an AI/ML Product Owner with a proven track record building scalable machine learning systems in regulated environments — currently at the U.S. Patent and Trademark Office (USPTO). My work focuses on transforming real-world workflows into ML pipelines and AI-powered services.
- 🚀 Lead end-to-end ML product development — from prototype to production
- 🔎 Build retrieval-augmented generation (RAG) systems for professional workflows
- 📊 Translate stakeholder needs into clean data pipelines running 24/7 in Production
- ☁️ Manage deployment of containerized inference services on AWS, Azure, and on-prem solutions
-
RAG-RAG Starter Kit
A reality-rooted RAG framework for federal compliance. No Digital DMT 🌀 included — this architecture keeps AI systems grounded in verifiable sources and compliant with federal requirements. -
CODE @USPTO Newsletter Project
Born out of laziness, a Python + React system to streamline newsletter creation and distribution for the Club for Open Data Enthusiasts (C.O.D.E.); not an official USPTO project :) -
Chat-MPEP
Indexed, embedded, and runs locally. Chat-MPEP transforms the USPTO’s Manual of Patent Examining Proce 59CD dure from thorny HTML format into structured JSON 🧼 and powers an interactive chatbot using LlamaIndex and Microsoft Phi. Demoed on an airgapped laptop at USPTO Community Day 2024. -
CPC Definition Expansion Tool
Sample code from a much larger project I am working on (to be released). THe goal is to create Human-readable definitions for 250,000 CPC symbols using LLMs and a deep respect for taxonomy. Finally a way through the classification rabbit hole without losing your head 👑