At the time, the biggest skill would be a solid grasp on Python and Unix/Linux fundamentals. I'm not sure if anything has changed since I left, but I do remember there being some discussion about using Swift to write new tools. In addition, it is key to have a good understanding of typical PC hardware and how the components relate, as the job pretty much entails writing software to measure the performance of new Mac hardware through various metrics, i.e. how fast to write and read certain file sizes to disk, frames per second of different settings of display resolution when running games, video editing software, or even just streaming Netflix. Although you'd be working alongside analysts who are ultimately responsible for interpreting the data, knowing statistics is crucial in creating reliable tests and accurate and useful workloads and sample sizes. Since there winds up being a huge number of tests to run every day on several different SKUs of Macs, you will need to be able to strike a balance between a quick enough test suite and statistically viable samples. You will also need to be good at troubleshooting and figuring out things on your own. With unreleased versions of macOS running on unreleased hardware, there are inevitable countless bugs. How do you get around those so you can get useful data to the engineers? It's an experience where you'll find yourself unable to count on Google or StackOverflow for assistance.