Screen with recruiter. Massive take home technical case, followed by a video interview where I presented on the case (the interviewer was not especially familiar with the case). Was told I'd be proceeding. That was followed by a frantic attempt by BCG to schedule more interviews for me (including an in-person). I gave them availability, and then they stopped replying to my emails. A couple of weeks later I got a form rejection. That stung a bit--the take home case was about 8 hours of work, and to not get thanked for my time after they selected another candidate (I presume, given the job posting is gone). The case is basically a Kaggle problem, except you're supposed to make a presentation for a mock client. The case is kind of silly -- it's a churn problem with standard modeling issues -- categorical data, missing values, noise, a poorly defined objective, a poor data dictionary, etc. The data appeared artificial, and modeled decently but only with non-linear techniques--it looked like they had created the data that way (there was a massive gap between performance of linear and non-linear models, far above what you would see in a real dataset). They also set some traps in the questions they asked you, where you really need to think about what's useful to the client rather than answering an academic problem.