Linear algebra is just a tool that fits nicely in a lot of computer science and various engineering toolboxes.
I'm talking super basic like when is it useful to transpose a dataset, to apply some function, then transpose back?
Being congnizant of the data-shape is something that comes as a side effect of studying linalg and matrices.
Just recognizing things like that will be the most value add unless you get into some domain specific requirements (games/engineering software).
You have a group of 20 people. Each of them rates how much they like everyone else, on a scale of 1 to 10. Your task is to figure out how popular everyone is, again on a scale from 1 to 10. Someone's "popularity" is based on how much the other kids like them, except the cool kids' opinion count more. You're popular if the popular people like you.
Can you mathematically compute how popular everyone is? Is the solution unique? Or might there be more than one solution? The problem seems hopelessly circular -- you have to figure out who is popular before you can figure out who is popular -- but actually it turns out to be a standard linear algebra problem.