Pros
The Kubrick model is a good programme to get grads into the data field. Having studied a non-STEM subject it would've been a lot harder without Kubrick to get into the field. The training programme is decent, nothing special/stuff you can't learn by yourself, but it is good having trainers you can ask questions to and motivate you to learn. To any potential applicants, it is a good opportunity that gets you good experience in the field and with decent pay but you need to make sure you are comfortable with a degree of uncertainty.
Cons
Uncertainty - You don't really get a say in where you end up (they give you a preference form to fill out but I don't think this gets paid much attention). A few of my training cohort got incredibly stressed over where they ended up. My advice is to go with the flow and try to accept what you get. With most places offering flexible working it isn't a big issue. You also get no say about what you'll be doing on client site which is fair enough but make sure you are comfortable with all parts of your job description because you could wind up using one piece of software and nothing else. There isn't much company culture (no office parties or anything) so if you're on client site without other Kubrick people you don't really feel like part of Kubrick.