Off-the-Shelf Solutions

In this course, we designed and built an ML training pipeline from scratch. While this is a legitimate endeavor, and several companies do build their own pipelines, it’s worth noting that there are some off-the-shelf solutions as well. Whether to use an off-the-shelf solution or build our own depends on the suitability of these solutions for the company and project, as well as other business requirements. Let’s take a look at a few free off-the-shelf solutions.

Kedro

Kedro is a framework that provides a scaffolding for ML projects. The scaffolding includes a standardized directory structure, configuration management, dataset management, data artifact versioning, model versioning, and logging. Kedro comes with the ability to visualize the pipeline. It’s also preconfigured for unit tests using pytest.