AI-Powered Real Estate Underwriting Platform MVP

AI-Powered Real Estate Underwriting Platform (MVP) PROJECT OVERVIEW: We're a commercial real estate investment/coaching business looking for an experienced full-stack developer to build the MVP of an AI-powered deal underwriting platform for internal use. THE CORE IDEA: a user file uploads a property's Offering Memorandum (OM), rent roll, and T12 (trailing 12-month financials). The platform uses AI to extract the relevant data points from those documents and auto-populates a structured underwriting model. The system then generates an AI-written analysis of the deal - red flags, positives, and due diligence questions to raise - based on that underwriting output. This is not a request for a landing page or a front-end mockup. We want a real, working application with actual document processing, actual AI-driven data extraction, and an actual underwriting engine behind it. PHASE 1 (MVP) - Keep It Simple We'd rather have something simple that genuinely works than something ambitious that doesn't. Phase 1 should focus on the core loop: 1. User uploads OM / rent roll / T12 (PDF, Excel, or similar). Examples can be given. They vary on format/data structure depending on each seller. 2. AI extracts key data points from those documents 3. Extracted data populates an underwriting model (income, expenses, NOI, debt service, returns) 4. User can review/edit the extracted data before it's finalized 5. System outputs a summary value/return analysis and a short AI-generated write-up of the deal (strengths, red flags, DD items). 6. Scalability to start with a “”high Level” quick and dirty analysis and later scale with a detailed full analysis. Nice-to-haves for a later phase (DO NOT scope these into the MVP price, just note them as PHASE 2 in your proposal): A saved deal pipeline/dashboard across multiple properties, notes/CRM functionality, and a business-plan builder. Non-Negotiable Requirements: Individual, private user logins are required from day one. This is the single most important technical requirement of this project. Each user's login should scope access to only their own deals and data no shared/open access. The underwriting logic must be customizable, not hard-coded - we have our own underwriting methodology/model that will serve as the basis, and we need to be able to adjust assumptions and formulas over time without a full rebuild. Because this tool will contain proprietary underwriting logic and confidential deal data, authentication should function as a real access barrier - not a cosmetic login screen. We do not want the underlying model or logic to be visible or extractable by someone without an account. Build with scalability and forward upgrade/expansion in mind. We don't expect enterprise-grade infrastructure on an MVP budget, but we don't want an architecture that has to be thrown away the moment we add a second feature. Please start proposal with "AI EXPERT".