AI based model trainer for text recognition (OCR)

Job Description

For our Digital Onboarding Toolkit product team we are looking for a capable AI based model trainer for text recognition (OCR). Someone who want growth in area AI, Machine learning.

Responsibilities

- We have designed an ML model for text recognition with a pipeline for its training.
- At start your main responsibility would be to prepare (configure, tune parameters) execute and post-process training for new documents (identity cards, passports of new editions and countries).
- Training image data generation and augmentation techniques (noise, blur, blending, etc.)
- Maintenance and improvements of existing python scripts for model training workflows
- Evaluation of accuracy / size / speed of trained models on datasets of documents (identity cards, passports, driving licenses, …)
- Possibility to work with neural networks

Candidate requirements

- Basics of Linux, Shell, Python, Git, Gitlab, CI & CD
- Basics of C/C++ is advantage
- previous experience with machine learning is not necessary
- motivation to learn machine learning and growth in this area

About DOT

Digital Onboarding Toolkit (DOT) is a full technology stack for developing remote identity verification applications. Consisting of mobile and server components, DOT can authenticate and extract data from identity documents and verify identity of an individual using industry leading facial biometrics and liveness check. DOT helps organizations and their users create instant trust with each other.