The way to write code that scales and accelerates your work as a knowledge scientist or machine learning engineer.

Once I was a junior Data Scientist, my goal was to write down code that simply worked.

I used to see Python as a framework to run Pandas, Numpy, or Matplotlib only. I began like everybody else in a Jupyter Notebook, processing the info and training models cell by cell.

I remember my first job in an organization.

Because the project progressed, the notebook grew, and despite providing explanations with markdowns, the code began to get messy.

The primary model was finally trained, its performance evaluated and shipped to production with the developers’ help.

Nonetheless, like all Machine Learning project, deploying a model is just not the tip of the journey but the start…

Several weeks later, I had to begin over and revisit the notebook. To be honest, it was almost easier to create a recent notebook. Requirements had modified. The code was too messy to try any modifications.

Moreover, shipping the processing algorithm to production was a painful task. Data needed to be processed identically across the notebook, within the training pipeline, and within the inference pipeline.

The necessity to write down the code 3 times meant that any modification within the notebook required corresponding changes in the several pipelines, increasing the likelihood of introducing bugs.

Doing Machine Learning at the moment was painful for me.

Until I began to apply Software Engineer best practices.

My code, my relationship with my colleagues, and my efficiency in delivering ML pipelines improved significantly.

One among those best practices was about using SOLID principles.

You almost certainly recognized yourself in my story.

Don’t worry—you’re not alone.