AI-Based Batch Document Processing System (with Manual Review Dashboard) Overview We are looking for an experienced AI / Backend Engineering team or freelancer to build a production-grade AI document processing system. The system will process multiple vendor document types in batch mode, extract structured data using AI, and expose APIs for downstream consumption. A web-based operations dashboard with manual review capability is required. This is not a PoC — we expect clean architecture, maintainable code, and enterprise-ready delivery. ⸻ Scope of Work 1. Document Ingestion • Batch ingestion of documents from Amazon S3 • Support for multiple document types (Invoices, POs, GRNs, Delivery Notes) • Scheduled batch execution (hourly/daily) • Error handling & retries ⸻ 2. AI-Based Processing • Document classification (vendor + type) • OCR and layout-aware field extraction • Vendor-specific field mapping • Confidence scoring per field • Routing low-confidence documents to manual review queue ⸻ 3. Data Storage • SQL database (PostgreSQL / MySQL) • PO-centric normalized schema • Audit logs and batch metadata ⸻ 4. APIs • REST APIs to fetch consolidated data by PO number • JSON response format • Pagination and basic filters ⸻ 5. Web Dashboard (Mandatory) • Processed documents queue • Manual review queue • Document viewer • Field-level editing & approval • Role-based access (Admin / Reviewer / Viewer) ⸻ Technical Preferences • Backend: Python / Node.js • Frontend: React / Next.js • Cloud: AWS • AI: OCR + LLM-based extraction • Processing: Batch only (no real-time) ⸻ Deliverables • Deployed solution • Source code & documentation • Database schema • API documentation • Deployment instructions Pricing Model Please propose your estimate under one or both of the following: 1. Client-hosted deployment (client pays cloud & AI usage) 2. Vendor-managed SaaS model (monthly pricing) Clearly mention: • One-time implementation cost • Estimated monthly operating cost • Assumptions (volume, vendors, fields) Proposal Requirements (Important) To be considered, please include: 1. Relevant experience with AI document processing 2. Sample projects or GitHub links 3. Proposed architecture (high-level) 4. Commercial estimate 5. Post-delivery support / AMC details