AI-Powered B2B SaaS Platform — Backend Build The React/TypeScript frontend is complete and exported from Lovable.dev to GitHub. The UI is in good shape — onboarding flows, admin portal, and module interfaces are all built. Some marketing content and copy will evolve during development but the frontend is not the focus of this engagement. I need an experienced full stack developer to build the backend, wire all integrations, and deliver a production-ready platform. Architecture is fully designed and documented. A detailed scope of works document is available after a brief screening conversation. I need someone to build to spec, not design the solution. The platform is a multi-tenant B2B SaaS product. Each customer organisation is a fully isolated tenant with their own users, uploaded documents, and AI interactions. The AI layer is grounded in tenant-specific uploaded documents via a RAG pipeline — meaning the AI responds based on what each business has uploaded, not generic internet knowledge. Staff interact through a chat interface. Every interaction is permanently logged in a full audit trail. A core trust requirement of the platform is data sovereignty and isolation. The platform reads Microsoft 365 data via Graph API but never copies or stores it in the platform database — M365 data stays in the customer's own Microsoft tenant at all times. All other customer data — uploaded documents, AI interaction logs, tickets, and audit records — must be stored in fully isolated per-tenant partitions in Supabase with row-level security enforced on every table. Zero data retention must be active on all Anthropic API calls — no query or response is retained by the AI provider. The developer must be able to articulate and implement this data architecture clearly — customers will ask about it and the platform's trust proposition depends on it. The core stack: Supabase — PostgreSQL, Edge Functions, Auth, Storage, pgvector Microsoft Graph API and Microsoft Intune — per-tenant Azure AD app registration model, reading M365 users and devices, executing device configuration policies programmatically. Each customer registers their own Azure app in their own tenant during onboarding via a guided wizard with pre-filled fields and an optional PowerShell script for automated registration. Customer app credentials stored per-tenant in Supabase Vault. This is the correct ISV architecture — do not apply if you have not implemented Microsoft Graph API integration before. Anthropic Claude API — RAG pipeline with pgvector, strict multi-tenant document isolation Stripe — subscription billing with metered add-ons and full webhook lifecycle Resend — transactional email with domain authentication Endpoint security provider integration — the platform must support a pluggable EDR webhook for device enrolment status and billing triggers. The specific provider will be confirmed prior to Phase 3. Build the integration layer to be provider-agnostic where possible. The onboarding flow must include an automated procurement notification triggered on device enrolment confirmation — a formatted order trigger sent to a configurable supplier email address with customer details and device count. The 14-day trial period is intentionally designed to allow licence provisioning to complete before billing commences. The work covers five phases: Credentials, infrastructure, and per-tenant Microsoft Azure app registration flow RAG pipeline with pgvector and tenant-isolated document retrieval Microsoft Graph API and Intune integration including automated device remediation with rollback Billing, user management, and onboarding flows Internal metrics dashboard, automated monitoring alerts, and additional platform features Non-negotiable requirements: Microsoft Graph API — direct hands-on experience required, not theoretical. Please do not apply without this. Supabase Edge Functions and RLS — must be comfortable building production-grade multi-tenant backends RAG pipeline with pgvector — tenant isolation is an absolute requirement Stripe subscriptions and webhooks Nice to have: Lovable.dev codebase experience Previous AI SaaS product builds Engagement: Fixed price, milestone-based payments across five phases. Full scope document shared after screening. Looking to start immediately. In your proposal please include: Specific evidence of Microsoft Graph API work Specific evidence of a RAG or pgvector implementation Your honest view on timeline for phases 1 to 3