Claude API Developer/AI Document Processing

Scope of Work Document Ingestion & Parsing • Accept uploads of PDF and Excel (.xlsx / .xls) commission statements from any aggregator • Use the Claude AI API to intelligently parse and extract structured data from documents, regardless of format differences between aggregators • Map and normalise fields across different aggregator layouts into a unified data schema • Handle 12–24 months of historical statements per broker client • Support batch ingestion of 5–15 client report sets per monthly processing cycle Core Analytics & Reporting The platform must produce the following analysis for each broker client’s loan book: • Settlement dates and loan age analysis (age distribution of loans in the book) • Average loan size — overall and segmented by lender • Lender concentration analysis (percentage of book held with each lender) • Trail income payments broken down by period and cumulative totals • Trail income trends and growth rates month-over-month and year-on-year • Run-off percentages over the reporting period (loans exiting the book) • Total loan book size over time, including growth and decline trends • Visualisations and charts for key metrics to support valuation discussions and buyer presentations Multi-Client Management • Ability to manage multiple broker clients, each with their own aggregator statements and reporting history • Per-client dashboard and analytics view • Strict data isolation between clients — broker data must never be visible across client accounts • Ability to flag or archive completed transactions Output & Export • Dashboard view showing key metrics and trends per broker client • Exportable reports (PDF and/or Excel) suitable for use in business sale presentations and buyer due diligence packages • Ability to compare reporting periods side by side • Summary ‘snapshot’ output that can be shared with prospective buyers as part of the Yield marketplace listing Technical Requirements • Must integrate with the Anthropic Claude AI API for document understanding and intelligent data extraction • Must handle both PDF and Excel input formats reliably across a variety of aggregator layouts • Web-based application accessible via browser — no local installation required • Secure login and encrypted data storage — broker and client data is commercially sensitive and must be protected • Scalable architecture to accommodate growth in broker client numbers and monthly statement volume • Clean, intuitive UI operable by non-technical business brokerage staff • Developer to advise on preferred tech stack and hosting/infrastructure approach Skills & Experience Required We are looking for a developer or small team with demonstrated experience in: • Integration with LLM APIs (Claude API by Anthropic strongly preferred; OpenAI API experience also relevant) • PDF and Excel document parsing and intelligent data extraction • Building data analytics, dashboards, and reporting applications • Web application development (Python, Node.js, or similar backend; React or similar frontend) • Experience in financial services, mortgage broking, or business brokerage software is advantageous • Strong communication skills — this is a collaborative engagement requiring regular updates and responsiveness What We Need in Your Proposal Please include the following in your response: • A brief overview of your relevant experience, including any similar document analysis or financial data projects • Your proposed approach to document parsing and handling format differences between aggregators • Recommended technology stack and hosting/infrastructure approach • Project timeline with key milestones and a proposed delivery structure • Pricing — fixed price, time and materials, or staged delivery with cost breakdown • Any questions or clarifications you require before providing a full scope