This repo contains the slides, notebooks and example code for the talk at code.talks 2018 in Hamburg.

You should install luigi before running the pipeline.

pip install -r requirements.txt

The example luigi pipeline is under "pipelines". It can be started by running

PYTHONPATH='.' luigi --module 00_training_pipeline Export --dataset-version 1 --model-version 1

The notebooks contain the tasks and a few examples for presentation purposes.

The slides are accessible via the Jupyter Notebook "Production-Ready-Datascience.ipynb" as well as the PDF "presentation.pdf".