'Shift Left' – running tests as often and as early as possible is the dream of QA and development.

But how, when my tests run for hours or days? When I have slow HIL tests or manual tests that are difficult to parallelize? The dream of an accelerated testing process in the embedded sector seems unattainable.

But there are solutions. Test selection methods choose a subset of tests that find the majority of bugs in the shortest amount of time, enabling significantly faster feedback.

In my presentation, I will introduce the foundation, research, and practical experience of two such methods.
One of them uses AI and clustering to create a quality gate that can find many bugs across the board in a short amount of time.
The other method uses test coverage to achieve a significant acceleration of testing even in particularly difficult environments, such as HIL testing with variants.