Best practices for building machine learning platforms on the cloud

Dmitry Spodarets
Dmitry Spodarets

Gartner predicts that by 2025, cloud platforms will underpin more than 95% of new digital initiatives - up from less than 40% in 2021. With cloud platforms that can handle large AI and ML workloads, companies are becoming increasingly efficient in providing a first-class customer experience.

Marcie Apelt - MVP, ML/AI Product at Capital One, based on his own experience leading enterprise technology platform teams, shared the best practices he's learned along the way.

An announcement of some of them.

  • Strong team. It should be a functional team of the best people, and it doesn't have to be big. At a minimum, the team should have product managers, engineers, and designers - people who understand the platform's users.
  • Doubling down on time. You have to continually estimate how long it will take to create and then double it. In Marcy's experience, that estimate ends up being surprisingly accurate.
  • Focusing on business outcomes. It's important to sequence the work in a way that achieves business value along the way. This motivates the team, builds trust, and creates a virtuous cycle.

Recall also that the industrialization of machine learning and applied AI were named by McKinsey as two of the 14 most significant technology trends of our time.