Why Cloud Config Mgmt is Mandatory for AI/Machine Learning

Eric Broda
7 min readSep 19, 2019

Cloud config mgmt issues are a big cause of data/security breaches and are stalling cloud AI/Machine Learning projects. Here’s how this can be addressed.

Photo by fabio on Unsplash

Towards More Secure AI/Machine Learning

Business demand for AI/ML is unrelenting. And to quench a commensurately voracious appetite for computing capacity, AI/ML’s march to the cloud has started.

Or so it seemed. The early adopters — those enterprises that now have some practical experience with data on the cloud — have been rocked by recent data breaches (for example, Capital One). Not surprisingly, this has put a halt to previously grand ambitions for wide scale enterprise use of cloud-native AI/ML.

Which has led enterprises to introduce new techniques such as identity-based security, zero trust, and zero leakage strategies to address a number of cybersecurity vulnerabilities issues.

Unfortunately, more needs to be done. In fact, it would appear that poor cloud configuration practices may be a major cause of many data breaches. What can be done?

This article describes how configuration management provides foundational capabilities that protect the sensitive data that AI/ML requires. And by adopting these capabilities enterprise’s will once

--

--

Eric Broda

I write at the intersection of Generative-AI, Data Mesh, Data Products, Data Marketplaces, and Digital Ecosystems