You have 0 further articles remaining this month. Join for free to read unlimited articles.


Over the past decade the development of ML and AI technologies has proliferated with unprecedented progress being made in many fields.

We now have AI systems that process, categorise and make predictions on data for ever increasing types of applications from various recommender systems, medical condition detection, teacher grading scores and many more. We have systems adding filters to photos, airbrushing images or creating new images from words. We have systems that handle our customer service queries, drive our cars and verify our identities. All this progress has resulted in the average automated system being significantly larger in scale and complexity than a decade ago. 

This acceleration is partially enabled and driven by the development of tools facilitating anyone and any company, from experts to novices, to develop and launch models to the world. In 2023 there were 1,426 companies offering ML tools, compared to 139 in 2012. This represents a +1,000% increase! Evidently there is a huge need for tools to support the various technical requirements of your ML workflows. However, simply adopting one of the sometimes costly tools available is not likely to be sufficient to solve your challenges and accelerate your ML development by itself. 

The cornerstone of your ability to stay competitive and to innovate with speed and confidence using AI and ML technologies is employing a holistic approach to your ML platform and operations. 

In this talk, I will discuss: 

  • What is ML platform and operations? 
  • How to assess your ML platform maturity? 
  • How to discover, understand and prioritise the requirements for your ML platform? 

Last but not least, I will discuss real lessons from building and scaling our own ML platform and operations from scratch.