A Founder and An image synthetic engineer, And working on VLM/OCR for improving the precision of models.
Synthetic data generation for hard cases in business documents.
imbalanced fields, table column merges, and small-text failures.
Designed to address data gaps that block OCR performance
when real samples are missing or expensive to label.
Making your edge cases become the assets to your business.
https://github.com/alrowilde/ocr-producer