I would say CS 180 is definitely one of the most useful and interesting classes I've ever taken at UCLA. HOWEVER, as for the professor himself, I would say he's not helpful at all. I stopped going to lectures after like 2nd or 3rd week bc his lectures suck. And since he strictly followed the textbook I just decided to study the textbook myself instead of wasting time attending lectures(btw the textbook is great as it provides you an overall skeleton of how different algorithmic paradigms work, and gives you abundant examples showing how to apply those different algorithms).
The workload is chill. You only have around 5 hw problems every week. However, some of the problems can be really hard, even impossible to do on ur own. But anyways there are solutions available online and it seems that as long as you write something on each problem you'll get 100...
The midterm was hard with average around 40%. The final on the other hand was much better(at least compared with the midterm).

Contrary to older reviews, Ostrovsky seems to have gotten better in terms of teaching. After a confusing beginning to class where he spent 2+ weeks talking about NP-completeness (a topic barely covered in other 180 lectures) as well as unclear lectures, he settled down quite a bit after week 4-ish and delivered lectures that are worth going to. This quarter, he decided to write on the blackboard instead of using slides, which has its pros and cons. A big pro is that he would talk slightly slower, but unfortunately he does not have the best handwriting, nor is he the clearest. Reading the textbook is essential in most cases.

Homework is definitely intended to challenge the students, but unfortunately most of them are basically un-doable without the solutions manual. That's not really a trait about his class though, as I hear this is the case with other professors too. (The textbook just has hard problems in general.) I'd still recommend doing them as much as possible on your own – and start early, as nothing beats the feeling of coming up with a solution yourself after thinking about it through over multiple days.

Exams are exponentially easier than the homework. Even so, this year the average for the midterm was in the 60s, though have actually gotten gradually easier over the past years. Most past midterms with Ostrovsky had averages in the 40s.

In general, not as horrible as other reviews make it out, but if you have the option of waiting one more quarter to get Sarrafzadeh, wait.

Ostrovsky is obviously a very accomplished, intelligent, and incredible computer scientist. His contributions to the field are immense—very obvious when he mentions an algorithm and says he once skied with its inventor.

However, he is an absolutely horrible lecturer.

Just yesterday, he asked us to review him nicely on course evaluations and on Bruinwalk because he spent a lot of time revamping this course to be more accessible. I appreciate that, and don't doubt that he spent a lot of time making this course better for us. He truly is a nice guy, very approachable, and cracks hilarious jokes to lift our spirits.

Unfortunately, I've got to say the truth—he isn't a good lecturer. He makes everything more confusing and spends too much time fussing about which marker works the best. He goes through at least 20 markers each lecture. Reading material online, on YouTube, and even the book taught me more about algorithms than him. Again, this is not a personal indictment on Ostrovsky—I love the guy, but teaching isn't for everyone.

Tecot screams incompetency. I went to his first discussion and left halfway. I could not bear listening to him scribble on his iPad with utter nonsense.

Instead, I started to (virtually, since I had a conflict) attend Levine's discussions. Absolute masterclass in discussions. Very approachable, nice, structured mini-lectures that filled in the gaps where Ostrovsky left to be desired. He should be the lecturer, not Ostrovsky in my opinion.

Ostrovsky is the worst professor I've had in UCLA. I am sure he is a distinguished researcher but he sucks at teaching. His lectures made me only more confused about the material that I had gone over with textbook and youtube. His lectures are disorganized and he ended up messing up our final letter grades only to change them a couple of days later because there was an error in his excel sheets!!! If you ended up taking this class with Ostrovsky be sure to skip his lectures and study the material from textbook and youtube channels.

Ostrovsky is not a great lecturer, but makes up for it by going through the course materials at a sluggish pace, usually only covering one or two topics in an entire lecture. I gave up on taking detailed notes after the first lecture, but I still showed up to every lecture to listen as concepts he discusses in class are fair game for exams so I would look up the algorithms/concepts mentioned on my own time and find better explanations online. Outside of that, the book is the best resource.

Tests/grades are going to be heavily dependent on the TAs for the course. I believe the quarter I took it we had the nicest TAs/grading in the history of the course and would give out full points for answers that may be slightly wrong on exams. Also gave 15 bonus points on the final by making it out of 115 points but only grading 100 points. Very very nice TAs, discussions are typically more helpful than lectures if you're confused but not mandatory to understand the class material. Every algorithm in this class has been explained to death online, save for a few of the ones later in the class, and you can find very visual/simple explanations for all of them online.

The material for this class is extremely dense and the the lectures are often extremely confusing as the professor isn't the best lecturer. However, he is extremely nice and patient and I even find myself often cracking up at his jokes and comments in class. He also gives a very, very solid curve and is very generous with the grading. He does record the class, but the quality of the recording is so bad I usually give up after 5 minutes.

The grade breakdown this quarter was:
HW 15%
Midterm 40%
Final 45%

Overall, the material of this class is rough, but it'll be ok in the end after the curve.

Ostrovsky is a really nice guy! The class in and of itself is pretty helpful to understand how to approach these algorithm problems, and I honestly enjoyed the content. In terms of workload, it's honestly pretty manageable-- the homework is reasonable, usually doable within an hour or a couple hours at most, and the midterm and final felt very fair to me (though the final was definitely tougher than the midterm).

In my opinion, the professor was a great guy. He did his best to issue PTEs to everyone who needed them, stayed after lectures to answer questions, gave multiple accommodations throughout the quarter-- allowing us to drop a HW, a separate grading scheme if you didn't do well on the midterm, and so on. You could tell from his lectures that he was really trying to ensure that he was going slower and explaining things more clearly so people could better understand.

HOWEVER. That being said, the lectures will still pretty tough to follow along. Partially because of the way that the material was presented and because of the sheer denseness of all the content too, it's just plain hard to follow. It doesn't help either that Ostrovsky doesn't always have the clearest handwriting or explanations. I'm more a visual learner, so for me I really had to push myself to strain and listen during lectures, and even then if I lost focus for just a minute I'd look at the board again and then be totally lost. I'm sure there were many students who were able to follow just fine, but if you find yourself in a similar situation, what I found helped me the best was just trying to write lecture notes and follow the best I could, but really just recording the "big topics" of the lecture. It's all in the book, so I'd then just reread the textbook and follow each example / proof closely so I fully grasped the concept. ChatGPT is great at explaining some of the more tricky algorithms too, so if you're ever lost on something, I'd recommend asking it for an explanation at a high level (when the book gets too formal).

Overall, interesting class, you can tell the professor means well, and I felt that the workload was reasonable. I would recommend taking this especially if you're more OK with self studying.

Ostrovsky’s lectures are often hit or miss. He is good at explaining some topics and awful at some others (such as NP-Completeness).

Make sure you read the formal proofs in the textbook and go over as many different problems as possible. This increases your chances of seeing a similar problem on the midterm/final (which was hard).

If you care about getting a good grade in CS 180, I'd say Ostrovsky is the best professor to take the class with. He gives the most amount of A's out of all the other professors. However, he is not a good lecturer. Personally, I don't care if a professor is good at teaching or not because there are plenty of online resources (Geeks for Geeks) and YouTube videos (Tushar Roy, Back to Back SWE, etc.) that explain the material much better than most professors can. I did close to 200 leetcode problems the summer before, so I had good algorithm knowledge before coming in, but it is not necessary to do well in the class as long as you study hard.

Don't even bother going to class. Just go the first week when he talks about the class, the class before the midterm, and the last week when he talks about the final. What I did most of the time was skip class and then ask my friend what was covered and then self study the material. It was often more efficient to learn on my own since Ostrovsky does not explain things well. Here's some of the material I used to do well in the class. Discussion were a hit or a miss. This quarter, all of the TAs were not good. Discussion is a hit or a miss, but usually a miss. Maybe in another quarter the TAs will be good.

Homework assignments are often free 100%. Don't even bother spending much time on these. Attempt the problems as best as you can, and if you can't get it, just copy the solution manual or some old homework on GitHub. I only got marked off on two assignments, but still got a 95/100 on those. One of the mistakes was not writing the runtime, so always write the runtime if it asks you to write a polynomial time solutions or something. Other than that, as long as you write something that sort of makes sense, you are good. The homework questions are often way harder than the exam questions, so it is not worth it to study the homework that hard. Even the TAs don't know how to do most of the homework questions. I often went to office hours to understand some of the homework problems and even they couldn't do them.

As far as his exams, I suggest doing most extra problems in the beginning of each chapter. As you go to the later numbered problems, they get much harder. Often, very convoluted and long questions won't show up on the exam. I found that Geeks for Geeks was the best way to prepare for the exams. A good amount of problems were pulled straight from Geeks for Geeks word for word. For the midterm, from what I heard, the TAs each picked one question to put on the midterm, and then Ostrovsky put one. The midterm was 6 questions. The final was 10 questions. I also found doing some extra problems in the CLRS textbook helpful. This CLRS textbook was more readable than the textbook we used in this class. The textbook we use in class is not that good in my opinion. I would just skim that book and then get a general idea on how to write proofs for each type of problem. The hardest part of the class was not really writing a proof for the questions on the exams. It was coming up with a solution. As long as you wrote some sort of proof that followed the format, it was usually good enough.

Ostrovsky gives PTEs to everyone, so you shouldn't worry about getting into his class, so don't even bother first passing this class. Everyone gets in, but people end up dropping anyways, so it doesn't really matter.

Here are some good resources to help you get through the class:
http://www.cs.princeton.edu/~wayne/cs423/

https://web.stanford.edu/class/archive/cs/cs161/cs161.1138/

https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-046j-design-and-analysis-of-algorithms-spring-2015/lecture-videos/

https://www.youtube.com/user/purpongie/playlists

https://www.youtube.com/user/mikeysambol/videos

Alright, Ostrovsky's lectures are getting better than what the reviews said, but I'd still read the textbook instead. The textbook explains everything in detail, and is pretty straight-forward except for the chapter with NP-Completeness.

Contrary to what other reviews say, I'd say the tip to doing well in this class is reading the textbook until you understand the material. (And proofs - you need to explain why your method is the most efficient). Solving more problems isn't necessarily the answer. I came in with no previous knowledge, just tried my best at understanding the material though the four HWs, and got an A. Please just don't copy the answers from online for the homework (you won't be prepared for the exam if you do!)

I personally feel like that the students who will do the best in this class would be Math majors who have experience with proofs. TAs are CS TAs so they don't fully expect full-math proofs. Just try to get your answer look like a Math-proof.

There are only 4 HW problems assigned per week but, it's gonna take quite some time. Many people suggest you to start early, but Ostrovsky often didn't cover the material until the day before the due date (at 8AM). I always ended up doing CS 180 homework at 2AM, though the TAs are pretty generous with grading the HW.

But to be honest, if you have a tight courseload, save yourself time by just reading the textbook and skipping the lecture. The only merit of going to his lecture is that Ostrovsky often hints what exam question will be on the exam (and might not be a freebie if you don't show up to class... Master's Theorem and Directed Acyclic Graphs for this year). I personally only showed up for lectures during Week 1 and Week 10, occasionally checking what the class is going over. Just ask a friend who's attending the lecture.

If you have a light courseload, I'd recommend you to read the textbook first, and just listen to Ostrovsky's lecture.

CS 180 really isn't a class where you can take good notes. It's more about understanding a concept and applying it. Proofs are the tools that help you understand why an algorithm works such way or why it's the most efficient.

And for Exams... It's really you see it or you don't.
You want to do some priming so you have a better chance of solving his problems. The exam questions however, aren't as difficult as the homework problems, and can be solved in given amount of time. (but your hand will hurt from writing a lot).

Read the example from textbooks as it helps you to understand the concept. Don't get too focused on proof of one example, since Ostrovsky's exams aren't about memorization. Try to learn how the process works. (Network-flow, Divide and Conquer, Dynamic programming)

He's a very kind professor though. Doesn't want many students stressed about his grades so his curve is always more lenient. I'd actually recommend taking CS180 with Ostrovsky if you are inclined to self-learn.

TL DR; Read the textbook (until you understand them). Kind professor with generous curve, but not the best lectures. Learn the concepts rather than memorizing examples.