Isn't there a large gap between the types of Fermi estimation problems you might get in a regular job interview (e.g. Microsoft PgM or McKinsey Associate) and the types of probability/statistical problems you might get in a quantitative finance interview?