You shouldn’t make “blanket judgments” that apply to every individual Bangladeshi. In a context where you’re evaluating individuals, you should do so without regard to their ethnicity or nationality. In the context where you’re talking about masses, then it’s permissible to draw inferences about aggregate characteristics.

Simple example: at the doctor’s office, I want them to measure my height, without regard to the fact that I’m Bangladeshi. But that doesn’t mean that medical journals must blind themselves to the fact that a group of 10,000 Bangladeshis is going to have a much lower average height than a group of 10,000 Germans.