
Example build
Supplier invoice extraction for an accounts-payable team (PDF to structured data)
- PDF and scan ingestion from a shared inbox and a watched folder
- Header and line-item extraction with per-field confidence scores
- Per-vendor templates that improve as invoices are approved
- Checks: line items vs subtotal, duplicate detection, PO match, GL account suggestion
Python, OCR, layout-aware PDF parsing, LLM extraction with validation rules, PostgreSQL, React
