At a glance
WHAT IT’S REALLY ABOUT
End-to-end app prototyping, deployment, and analytics using Claude on GCP
- The presenter demonstrates using Claude Code to turn a rough sketch into a wireframe and automatically open a GitHub pull request, even for non-technical roles like a PM.
- Using Claude Code’s planning mode, the wireframe is converted into production-style UI pages and merged via PRs to simulate a real product workflow.
- Google Cloud’s Developer Knowledge API (with an MCP server) and Google Cloud “skills” are used to design and implement a Cloud Run + Firestore + BigQuery + Looker architecture with up-to-date documentation context.
- Claude Code subagents parallelize implementation (API, data pipeline, dashboard) and generate CI/CD using Cloud Build and Cloud Deploy to ship to dev and then promote to production.
- A custom security-review plugin adds input validation and tightens service-account permissions, after which data is analyzed via BigQuery MCP and visualized via a Looker dashboard for product insights.
IDEAS WORTH REMEMBERING
5 ideasClaude Code can compress the PM-to-prototype cycle into minutes.
A hand-drawn concept can be converted into a usable wireframe and committed via an automated PR, reducing handoffs and waiting time for early UX iteration.
Planning mode is a practical guardrail for UI build quality.
By forcing a spec-first step before code generation, teams can align on page structure and requirements (e.g., landing/feedback/thank-you/admin dashboard) before implementation begins.
MCP-based documentation access reduces “I don’t know GCP” friction.
The Developer Knowledge API’s MCP server gives Claude fresh, structured guidance (updated daily) so architecture and implementation decisions are grounded in current platform docs.
Subagents make AI-assisted development feel like a small sprint team.
Splitting work into API, data pipeline, and dashboard subagents enables parallel progress and faster integration, with tests run after implementation to catch regressions early.
CI/CD can be generated as part of the app—not bolted on later.
The demo shows Claude producing Cloud Build (CI) and Cloud Deploy (CD) configuration so merges trigger builds, releases, and dev deployments automatically.
WORDS WORTH SAVING
5 quotesSo how many of you used an AI coding tool this week? Raise your hand.
— Ivan Nardini
But with respect to this team, Claude Code, uh, the Anthropic's, uh, coding agent, provides a set of capabilities that will essentially augment them, uh, across this entire, uh, software lifecycle.
— Ivan Nardini
So today, what I'm gonna do, I'm gonna put on, uh, I'm gonna put on five different hats, and I will show you how you can leverage Claude run, uh, Claude models running on Google Cloud, uh, to build and deploy a simple, uh, feedback app, uh, that it will be used at the end of this, uh, session to provide me, uh, a feedback on, uh, my performance here on the stage.
— Ivan Nardini
So first of all, if you use Claude models on Google Cloud, you pay per token.
— Ivan Nardini
And again, even if you don't know, like, a dashboard tool, you were able to build this using, uh, Claude Code and, uh, the MCP server that we provide.
— Ivan Nardini
High quality AI-generated summary created from speaker-labeled transcript.
