Without the need to use hyperscalers, we would like to create our own machine learning model that helps us recognize the pattern of about 10 different ID documents and allows us to read the ID information.
The OCR software to be used is Tesseract, the MLM is to help identify the areas read out so we know what is first name, last name, date of birth, expiration date, etc. It should also be possible to detect rotated documents so that Tesserat can read them.
It is important to have knowledge for Tesseract and to be able to evaluate the hosting conditions necessary to implement this solution cloud-side. So it is needed to be able to think a bit outside the box to adjacent areas and not only to focus on MLM.
This job is already closed and no longer accepting applicants, sorry.
