Just wait until you start interviewing at even crazier companies than FB/Google that would like you to tell them 20 ML algorithms, 20 computer vision algorithms, compare expected results of non-linear optimizations applied to some weird Deep Learning problem and analytically justify it (why is it worse/better than SGD on this loss function?), then off to some deep domain knowledge, conjure graph algorithms on the spot on problems you've never seen before, all in 1.5h, expecting Stanford/MIT PhD-level answers.