GDG Tulsa will host an in-person build session for Google's Gemini Live Agent Challenge on March 16, designed to take builders from zero to a working submission. The session walks the full hackathon spec end-to-end: getting access to the Gemini Live API, configuring streaming audio and video, building multimodal agent behaviors, handling tool use and function calling, and packaging the project for submission. We cover the official judging criteria, scoping tips for solo vs. team builds, common architecture choices (React + Vite front-end, Cloud Run back-end, Cloud Storage for assets), and deployment paths through AI Studio, Google Cloud Run, and Vertex AI. After the walkthrough, the room shifts into open build time — Q&A, troubleshooting, and live debugging with anyone working on a submission. Attendees leave with a working starter, a deployment plan, and a clear path to a qualifying entry before the challenge deadline.