Titanic Dataset overfitting: can it be that much?

Question

I am a bit confused as I am training a model that yields circa 88% CV score on the train data while the same model performs poorly on the test data after I submit it ( score