I am ready to turn my proof-of-concept into a very lean web-based MVP that showcases how AI can streamline credit underwriting from raw bank statements all the way to clear-cut risk insights. The goal is to let a lender log in, drop a statement, and instantly see a visual, data-driven picture of the applicant’s cash-flow health and risk profile. Core scope • Bank statement upload: accept PDF, CSV, and common image formats, then parse them server-side. • Automated data analysis: run my existing Python model (or one you help refine) to categorise transactions, calculate key ratios, and flag anomalies. • Risk insights & financial dashboards: present cash-flow trends, concentration risks, and a headline risk score via a responsive, interactive UI. Tech notes I prefer a standard Python backend (Flask or Django) with an API layer that separates data ingestion from the React/Next.js front end, but I am open if you can justify a different stack. Please keep the architecture clean enough for quick iterations and easy hand-off. Acceptance criteria 1. A hosted URL where I can upload sample statements and receive analytics within 30 seconds. 2. Dashboard renders in the latest Chrome and mobile Safari without layout breaks. 3. Parsed data and generated metrics are stored in a simple relational database, exportable to CSV. 4. Clear README with setup steps, model entry points, and a short Loom or GIF demo. If this sounds like a build you can execute rapidly and cleanly, tell me about one similar data-heavy web app you have delivered and your estimated timeline to hit the acceptance checks above.