The Good: Prof. Reidl is very good in explaining, going over his lectures twice really has a high coefficient of knowledge retainment, you will be ready for Exam1.

The Bad: Maybe it applies only for the Summer25 course, but the final exam was prone of TA"s subjectivity and required muscle memory. There were many students in Discord complaining that NLP being the last course, for them it was the hardest exam.

The Ugly: Not unique to this course, but some lectures given by some "professors" or TA's are really embarrassing for any university. Their English is incomprehensible, never mind their pedagogy (for everyone who took this course knows I am talking about META videos.)

Difficulty: The most confusing in evaluating this course is "the difficulty" assessment for this course. If one wants to understand what's going on in it, e.g. with transformers, it's very difficult, especially if you have no experience with pytorch. There is no way a student can learn pytorch and finish the assignments on the fly if not for the TA's and the professors making that "easy". So the professor has made it easy to write out the skeletons for us . People turn into AI for better understanding and as expected more than a few students got referred to OSI, even more were filtered as "cheaters" but they were not referred to OSI due to being hard to prove it. So, not sure where the "easy" part was in this course if things are fair and square. For the first time on this course, Exam-1 was on multi-option or single option questions. I personally did good (~97%) but I heard a lot of people bombed it. The final was one of the hardest tests I ever have taken in OMSCS, writing essays on some questions. Now, like everywhere else, essays bring a lot of subjectivity, a good portion of us thought the grading was lenient and some of us thought was very tough. There were a lot of students who couldn't get an A, but majority were not prepared for such workload considering this was supposed to be an easy course.