How I Built an AI Tool That Processed 15,000 Prescriptions in 48 Hours

When I first heard about the charity healthcare event, the challenge seemed insurmountable. Over 15,000 handwritten prescriptions needed to be digitized into structured JSON for pharmacy management in less than 48 hours.

The Architecture

I built OmmSai using a combination of Python, the Claude 3.5 Sonnet API, and a highly concurrent processing engine. The core of the system was a multi-threaded batch processor that could handle hundreds of requests per minute while maintaining strict accuracy checks.

async def process_prescription(image_bytes):
    # Convert image to optimized base64
    # Call Claude 3.5 Sonnet with vision
    # Validate structured JSON response
    # Save to Supabase

The results were staggering. We achieved a 99% accuracy rate, significantly higher than human data entry, and completed the entire dataset with hours to spare.