Cluster 4 - MLOps tutorial for automatic coding in NSIs

Goal

This repository aims to provide a step-by-step guide to building an MLOps pipeline. It presents the main building blocks of an ideal MLOps pipeline, together with feedback and example implementations from three NSIs (Insee, German Federal Statistical Office, and Statistics Austria) and other project partners.

This work was carried out within Work Package 10 "Text-to-Code" (WP10) of the AIMLL4OS project, and more specifically within the Cluster 4.

WP10 aims to explore and apply AI/ML methodologies to enhance the accuracy and efficiency of classification and coding processes used by National Statistical Institutes (NSIs).

You can find more information about WP10 on the CROS website, its GitHub Repository and its dedicated GitHub Pages.

Tutorial Overview

(Still WIP)

Code Structure

(Still WIP)

Runnable Example

This repository follows the AIML4OS template provided by the Work Package 6.

It demonstrates the work carried out within WP10 and its cluster by combining documentation and short code example.

Useful Documentation and Links

You can access the MLOps Tutorial here