Develop AI-Enabled OCR Prototype to Digitize Caregiver Records

Notify Health is looking for a volunteer to help us build and test a prototype OCR/AI system to automatically extract and digitize caregiver contact information from photos of paper health records.

Currently, we manually digitize these records to send automated immunization reminders in Nigeria. Automating this process could save over $100,000 annually and allow us to reach tens of thousands more children.

Deliverables for this project include:
1. Evaluation of existing OCR/AI tools (e.g., AWS Textract, Google Vision, or Tesseract) against a sample set of our health record photos.
2. Development of a functional prototype/script that extracts specific fields (names and phone numbers) from these records with a measurable accuracy rate.
3. A brief technical recommendation report on how to integrate this prototype into our existing workflow for scaling.

Notify Health protects children in Nigeria from preventable illness and death by delivering automated voice and text immunisation reminders to their caregivers. To maximise immunisation protection during a child’s vulnerable first year of life, we gather caregivers’ contact information by digitising paper health records as soon as possible after their child is born. At scale, automating and improving the accuracy of the photo review and digitisation processes could save over $100,000 annually in digitisation costs and enable us to reach tens of thousands more children each year.

We have a documented manual workflow and a dataset of sample photos ready for testing. We will hold regular check-ins to iterate on the prototype. We are looking for a volunteer with experience in Python, OCR, and Machine Learning.