I started out spending closer to 10 hours a week earlier on in the AI course. We initially had a local meetup and I spent more time reading the textbook for the first few units. Had to scale it back to 5 or so hours, which is good enough to get through it, but not enough to spend extra time discussing with others, reading supplementary materials, and coding.
For me it took 3 or 4 hours a week per course. It doesn't sound like much, but pre-arranged plans made it hard to find a decent sized piece of time (>1 hr) in which to do the work.
Thanks. This is the critical nugget of information missing from the course information pages (unless I'm just lacking in observation skills).
For me it was maybe 5-6 hours per week (for AI course).
It was quite intensive thinking, though, really requiring concentration and effort; the courses themselves (ie just video) were maybe about 1 hour per unit, and there were two units per week.
I've put about 2-4 hours a week into the ML class. However, I know that math (basic linear algebra and calculus so far) and have done scientific computing before so I can churn the exercises out pretty quick. The exercises are the best part. It's nice having a set of problems to work on instead of just watching an hours worth of videos each week.