At Broadstone Technologies, I designed a proof-of-concept AI architecture for MamaTrack AI, connecting the mobile app to LLM services through an MCP server and Supabase tools. I also developed a RAG-based approach for pregnancy content and structured tool-calling flows for application data.
I created and tested a Python MCP server using FastMCP and Supabase, and planned a modular design with Claude as the initial model and room for other LLM providers.
At Broadstone Technologies, I developed SpendLens / VESGO, an AI-powered finance application that parses PDF and CSV bank statements, categorizes transactions, detects subscriptions, and presents budgeting insights. I also built production features for Veilix, including WireGuard integration, dark-web monitoring UI, and cybersecurity learning features.
Previously, as Android Team Lead at Horizam, I led a cross-platform team of 7–8 engineers and owned architecture decisions and release delivery. I also shipped NFC-based digital networking capabilities for Vlinq, including tap-to-share, QR generation, Apple Wallet/PassKit, and lead management.
