Document AI / OCR
Also written as OCR, Optical Character Recognition, Intelligent Document Processing, Document Understanding
Turning scanned or photographed documents such as invoices, contracts, forms and IDs into structured data a system can use. OCR reads the text; modern document AI, often using multimodal models, also understands layout, tables and which value belongs to which field.
Think of it like
An assistant who not only types up a paper form but knows which box is the invoice total and which is the date.
Junior or senior?
Most real work is in edge cases: poor scans, handwriting, unusual layouts.
Senior sounds like
Can quote accuracy on real documents and describe how low-confidence results got sent to a human.
Ask them
“What was your extraction accuracy on real documents, and what happened to the ones the system wasn't sure about?”