document ai

A convincing field score can conceal a wrong purchase order. Define the complete unit of work, preserve page and row evidence, and calculate the review queue before choosing autonomy.

A design trade-off with cost, accuracy and latency all pulling differently. The interviewer wants a measured comparison on real receipts and a hybrid that sends each image down the cheapest path that works.

A regulated pipeline where the model reads documents but must not be the one deciding who gets an account. The interviewer wants extraction, checks against authoritative sources, fraud signals, human review and data minimisation.