Restaurant OCR Automation

Worked with a multi-location restaurant chain to replace slow, paper-based operations with a fully automated AI and OCR-powered management system. The business was handling thousands of handwritten delivery notes, supplier invoices, kitchen forms, and staff summaries every month, making it difficult to track food costs, inventory, and supplier spending in real time.

I designed and developed an intelligent document processing pipeline that allowed staff to simply scan or upload documents from their phones. Using Google Vision OCR, GPT-4, and a custom AI validation layer, the system could accurately read handwritten notes, printed invoices, mixed document formats, and low-quality scans, then convert them into structured operational data within seconds.

All extracted information was automatically pushed into a centralized live dashboard where management could monitor inventory movement, supplier costs, food cost trends, account statements, and shift summaries across every restaurant location in real time. The platform processed over 1,000 documents per month with 97%+ extraction accuracy while completely eliminating manual data entry.

Tech Stack: Python, FastAPI, PostgreSQL, React.js, Node.js, Google Vision OCR, OpenAI GPT-4, REST APIs, Cloud Storage.