EthioTrust – Distributed Backend & AI Vetting (Lead)
Developed an automated AI vetting pipeline to validate NGO documents, using OCR-derived text to determine document validity and reduce manual review. This work involved extracting text from document scans and applying rule/model logic for classification-style validation outcomes. The solution was designed to run at scale with asynchronous task processing to support high-volume document checks. • OCR using Tesseract to convert document images to text • Automated document validation/classification pipeline • Integration of OCR outputs into a scalable decision workflow • Deployment of the processing stack on AWS for reliable execution