At Google, I built the retrieval and data layer behind a Gemini-based onboarding assistant, indexing 12,000+ documents across 8 internal systems. I modeled employee and organizational context so answers reflected each person’s role, location, and employment type.
At Zayo Group, I developed an AI-powered recommendation platform that used Claude, LangGraph, RAG, and vector search to turn customer requirements into validated solutions. I also cut quote generation from 40 seconds to under 2 through batching, calculation optimization, and caching.
At Capital One, I owned the data layer for the Training Assignment Platform, consolidating learning data and supporting course recommendations. I rebuilt bulk assignments around batched requests and staged loads, reducing a 5,000-person learning plan from 6 hours to 20 minutes.
At eBay, I built the data pipeline behind personalized offers, processing 2B+ daily behavioral events and writing 180M+ user profiles into eBay Vector Similarity Engine. My work draws on a deep background in backend systems, distributed data infrastructure, and production AI applications.
