From the course: Securing the AI/ML Development Lifecycle: A Practical Guide to Secure AI Engineering
Join today to access over 25,500 courses taught by industry experts.
From DevOps to MLOps
From the course: Securing the AI/ML Development Lifecycle: A Practical Guide to Secure AI Engineering
From DevOps to MLOps
- [Instructor] Now that you understand why AI and, in particular, machine learning requires different processes, let's look in detail at what those processes are. For many in security, this is going to be something new that we haven't seen before. The important thing to keep in mind is that some of these activities don't fold into existing application security efforts, and assuming they do actually creates risk. As one example, most dev processes include validation of supporting components like libraries and APIs. They might be scanned for vulnerabilities so as to not accidentally introduce back doors. But did you know training data can introduce back doors too when that data is used to train an AI? Existing review processes won't catch something like this. This means that we need to adapt those processes so they do. To do that, we need to understand the specific steps and phases involved in building an AI product. Because…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.