Codebase supporting my talk on CI/CD for MachineLearnia (Nov 12 2021)

The dataset used is available here.

The point of the talk is to demonstrate a simple end-to-end data project management, with code linting/quality, models training and monitoring, CI/CD and automatic deployment on Azure Webapp.

The tools used are:

• MLFlow (Model training and monitoring, experiments and runs)

• Poetry with following packages:

  • flake8 (pep8 standards)
  • isort (imports sorting)
  • pytest (code testing)
  • bandit (security checking)
  • safety (at-risk libs checking)

• Github Actions (CI / CD)

• Flask (web app creation)

• Azure Machine Learning (Machine Learning environment)