At MAQ Software, I built the data ingestion layer for a Python RAG pipeline, preparing structured and unstructured data for vector search. I also improved a semantic-search system that turned natural-language queries into directory-path actions, reducing average response time from 45 seconds to 25 seconds.
I developed multi-step AI agent workflows with Microsoft Semantic Kernel and implemented Azure Functions for insight generation. I also connected a C/.NET backend with a React frontend and contributed to a Playwright and TypeScript automation framework for UI, API, and desktop workflows.

