Backend audit & launch-readiness plan — Flutter AI assistant on Firebase + Vertex + Cloud Run

What we are An early-stage AI first app. The Flutter/Dart front end is in good shape. The back end is a Node service on Google Cloud Run talking to Firebase Auth, Firestore, Vertex AI (Gemini), Google Cloud TTS, Google OAuth (Drive/Calendar), and Stripe. There is an existing internal audit, a terminology dictionary, and a documented data model. What we already know Voice (listen + respond) works end-to-end on the Chat overlay path Vertex AI + ADC is the chat rail; Firestore drawers + Cloud Run engines work Front-end voice playback, transcript, and TTS are wired An internal audit (Sessions 0–13) flagged ~20 backend gaps across Firestore rules/indexes, env config, OAuth tokens, Stripe webhooks, queue routing, and the factory/first-run path What we need from you A short, focused paid audit and plan that tells us exactly how far we are from one specific finish line: Start line = a closed-tester APK in Firebase App Tester with ~100 users, feedback piped to a Discord/WhatsApp channel. Not the App Store. Not paid. Just working, observable, safe MVP testing. You will be working alongside our Flutter lead. He will tell you what the front end expects from the back end. You will tell him what is real, what is half-wired, and what is broken on the back end. Deliverables Backend reality report (markdown, ~5–10 pages): Confirm/refute our internal audit findings on Firestore, Vertex, Cloud Run env, Stripe, OAuth, Functions deploys Add anything we missed Group items as: works · wired-but-unverified · broken · out-of-scope Path-to-tester plan with a numbered backlog: Each item: one-line problem, fix approach, effort (S/M/L), order Total item count is the answer to "how far are we from the start line?" 30-60 minute Google Meet video walkthrough of the report + plan (Meet/Drive) A short list of risks that could break testers in week 1 — ranked What we will give you Repo read access (Git + docs folder; ~270 docs including framework + audit) Access to Firebase Console, GCP Cloud Run, Stripe Dashboard (read-only first) Direct Slack/Discord channel with our Flutter lead A 30-minute kickoff call Hard requirements Strong Firebase + Firestore (security rules, indexes, deploy topology) Strong Node.js on Cloud Run / GCP (env, secrets, logs, IAM, ADC) Hands-on with Vertex AI / Gemini Stripe subscription + webhook experience Google OAuth (Drive/Calendar scope flows) Comfortable with AI coding assistants — Cursor / Codex / Claude Code / Continue / similar. Vibe coding and agentic workflows are part of the tripod. You should be the kind of person who treats agents as a tool to move faster, not a replacement for judgment. Plain-English written communication — we will read your audit Nice to have Experience working in/with founding teams Experience architecting, leading and building production grade products Experience in backend management of apps with large userbase An understanding of building with scale in mind How to apply Send a short proposal (no AI-generated boilerplate) with: One paragraph: why this project, why you, why not anyone else. Your closest prior project (link or description) One question about our stack you'd want answered in the kickoff What you need to know to give a timeline you can stick to. We will reply within 3 days. Two interviews max. One 30-min audit kickoff if selected.