She doesn't attribute malice to anyone. She describes the belief that the issue with models in the meltdown was inaccuracy as "maliciously" wrong, which I take to be hyperbole on her part to emphasize the category flaw she sees in this (namely, that if it's just inaccurate models, then better mathematicians are what's needed; in her view, the financial system at the heart of the meltdown was "corrupt and criminally fraudulent", so it's not mathematicians that are needed, and concentrating on them allows criminals to go free).

The fundamental problem she sees is that incentives in the financial industry are not aligned with accurate models, that inaccurate models were deliberately used to further short-term performance at the expense of further inflating the bubble.

"most of the misallocation of capital was done by firms doing things like trusting AAA rated securities"

On this, she links to specific evidence that the "trust" they were exercising was knowingly misplaced because the incentive was always to get the next big bonus.