Building a Machine Learning platform – let’s talk

Get a data engineer’s perspective on how effective collaboration is vital at each stage of developing our machine learning (ML) platform.


When Adevinta started working on a Machine Learning (ML) platform, we decided to internalise the process and create a platform that accommodates our specific needs. Read the full Medium article to get a data engineer’s perspective on what we needed from an ML platform, why collaboration was vital and how our teams tackled the development process.

Highlights include:

  • Identifying the steps of the ML pipeline
  • Issues around user experience
  • Principles to describe how the ML platform and its development should work
  • The advantages of making Kubeflow integration an iterative process
  • The type of spaces we created in order to share the development of the platform