Senior Full-Stack Engineer — AI SaaS Rebuild (Next.js + GCP + Multi-Model AI)

Senior Full-Stack Engineer needed to set up the technical foundation for a new web app. This is the first phase of a bigger build. I'm keeping it as a contained, fixed-scope project to start clean and make sure we work well together before scaling up. If this phase goes well, you'd continue with me on the rest of the build. You won't be inheriting any old code. It's a fresh start on a modern stack. What I need built in this phase is the backend and infrastructure foundation only. App features aren't part of this phase and will come later, so I'm not going into those here. Here's the work: Set up a clean Next.js project (App Router) with TypeScript and a sensible folder structure. Provision and configure the GCP infrastructure: Cloud Run for deployment, Cloud SQL (PostgreSQL), Cloud Storage, and Memorystore (Redis). Set up environment and secrets handling properly. Implement authentication with Firebase Auth (email/password plus OAuth), with session handling and a clean user model. Design the initial database schema and data-access layer (I'll give you the entities) and set up migrations. Build the AI routing layer. This is the part I care most about. I need a provider-agnostic layer that routes requests to more than one AI provider (Vertex AI, Anthropic, and others) behind one internal interface. Models need to be swappable through config, and it has to fall back automatically if a provider fails or times out. Streaming support and proper error handling included. Containerized deploy to Cloud Run, a staging environment, and a basic CI pipeline. A short README covering setup, the main architecture decisions, and how to run and deploy it. What I'll give you: the data entities for the schema, project-scoped API keys and provider accounts, and my architecture preferences (I have a clear target stack in mind). Stack: Next.js (App Router), React, TypeScript, Node.js, Tailwind, GCP (Cloud Run, Cloud SQL / PostgreSQL, Cloud Storage, Memorystore / Redis), Firebase Auth, and multi-provider AI through a custom abstraction layer. Who I'm looking for: Senior full-stack, 3+ years with Next.js, React and Node. Real hands-on GCP experience (Cloud Run, Cloud SQL, Cloud Storage). You've integrated more than one AI provider before and understand why an abstraction layer matters. You're comfortable with streaming responses, retries and fallback. You write clean, documented code, not just code that runs. Good communicator, works async, solid written English. This is fixed price, paid across milestones. In your proposal, suggest how you'd break the work into milestones within the budget. How you structure it tells me a lot about how you think. Again, this is the first phase of a larger ongoing build, so if it goes well there's a lot more work. Please don't apply if you don't have real GCP and multi-provider AI experience. A few questions I'd like answered in your proposal: 1. Have you deployed a Next.js app to Cloud Run? Tell me briefly how you set it up, how you handled environment variables and secrets, and which database you connected. 2. Have you integrated more than one AI provider in the same app (say Vertex AI plus Anthropic or OpenAI)? How did you structure things so the models were swappable, and how did you handle a provider going down or timing out? 3. When you start a brand-new production codebase, how do you keep it clean and scalable as features get added later? Walk me through your conventions.