Fall 2016 - This was a challenging class without having taken some of the lower divs that offer a primer on some of the topics. The class moved fairly quickly. Homeworks generally take a long time if you are doing them by yourself. I thought it was necessary to go to lecture because the notes helped a lot on the problem sets. Her lectures follow a logical progression through the topics. Overall a positive experience I definitely learned something in the process.
Winter 2019 - Honestly... she's probably the best you're going to get for 113. Her curve is set so that median is a high B+, pushing to an A-. But the way she runs her classes is problematic. Her examples and lecture notes are too simple, and are designed to teach you definitions and to present formulas to solve questions with (e.g. determining if a system is linear, TI, casual, stable). The homeworks and discussion sections are designed with similar principles, but much harder. Then you realize the goddamn midterms are EVEN HARDER, but on top of that, a large portion will usually involve proving some crazyass property and the questions related to homework and discussions make up maybe only 30% of the exam. And to nobody's surprise, the class got absolutely dumpstered. So on the final, the focus was on stuff similar to homework questions, which is good! But the test is now 36. FUCKING. PAGES. Just endless questions (i.e. endless torture). So that's already bad design. But on top of that, because of her grading scheme, and the fact that the midterm averages were so low and unrelated to the design of her final exam, it made me realize that I totally didn't need to even take any of the midterms. That's not a good outlook on the class, just fyi. Anyway she has an extra credit project. Do it lol.
Fall 2021 - This class, and the prof frankly, were a mess from start to finish. The book that the entire class is based on is from the 70's and is horrible at explaining the concepts taught in class, and the professor does the same along with it. Expectations for the class are extremely vague, with lessons barely correlating with any of the exams or homework. Additionally, there are two group projects that you are required to work on, and the first project was horribly explained and it took over a week to get clarification answers, with the project taking an excessive amount of man-hours with no clarity on how grading worked. Finally, as stated earlier, the homework and exams had almost nothing to do with what was taught in class, often asking very obscure questions that would often take much longer to decipher than to answer. Summary: The book, lectures, assignments, and the class as a whole are extremely unorganized and I would recommend avoiding this class unless you have a good understanding of graph theroy previously
Fall 2023 - I enrolled in this class as a CS undergrad. This class is hard, but very manageable; Fragouli does a great job at breaking down the subject and the course notes are very comprehensive. You do not need to be very good at linear algebra to succeed in this class (Math 33A is probably enough). You should be comfortable representing systems of equations as matrices and taking transposes of block matrices. Although nothing fancy is required in terms of linear algebra, this class is still math-heavy and mostly theoretical. The homework can get difficult, but prepares you well for exams. If you invest enough time, you will do well. The TAs and discussion sections are also very helpful. There was a group project this quarter (up to 4 people) about applying linear programming to machine learning; you will have to write code and do a report. The latter half of the class has an algorithmic focus, especially when it comes to Max Flow and its many variants. CS 180 will help, but not by much.