At Naya, I built Naya Studio’s core from the first commit, including its multiplayer canvas and live-collaboration layer. I shipped 120+ production releases.
I developed the custom whiteboard canvas and the client-server collaboration system for presence, cursor sync, reconnection, and keep-alive. I also owned authentication and access control.
For Naya’s AI cost-estimation products, I led prompt and rule-pack design, frontend delivery, and validation for enterprise clients including Miller Knoll and Schneider Electric. The apparel rule pack reached 98.9% material-cost match to an expert cost sheet with 0% run-to-run variance.
For the Schneider Electric should-cost benchmark, I iterated prompts and rules against reference teardown data, bringing mean cost error from 53% to about 18%. Before Naya, I built AngularJS components and configurable Jupyter Notebook interfaces for Deutsche Bank’s internal credit-risk tool.
