Algorithmic Toolbox

One of the best Computer Science algorithm courses (and hopefully, entire specialization) on Coursera's new platform. Here's why:
- The course supports programming assignments in multiple languages: C, C++, Python, Java. You can implement your algorithms in all 4 languages and learn all of them. They have automatic grader for all 4 languages.
- Your algorithms need to be optimized to pass the assignments, not just creating output correctly. The grader was designed to test you on 3 criteria: Correct Answer, Time Limit and Memory Limit. This course really forces learners to implement the best algorithms possible, not just a working algorithm.
- The materials presented are very well-designed. You can tell that tons of efforts have been put into developing the videos, the slides, the assignments.
- Talking about the assignments, each week there is one problem set that consists of 4 or 5 smaller problems. These problems range from easy (discussed in lectures, you only need to write code) - medium (not hard, but you need to design your own algorithm) - advanced (hard, and you'll need to design your own algorithm. They are not trivial. They are really challenging, and you'll spend a lot of time doing them, especially the advanced problems, but you'll learn a lot of things in the process.
- About algorithms itself, this course introduces recursion, naive and efficient algorithms in the first lesson. Next, it teaches greedy algorithms, divide-and-conquer and last but not least: dynamic programming. For each topic, there is one problem set consisting of 4-5 problems ranging from easy to advanced.
- The discussion forums are so lively with people constantly discussing correct solutions, different approaches to the problem, algorithms efficiency, even differences between languages! I have learnt so much from the forum, note that it is just the first week. The discussion thread for Programming assignment 1 alone, in just a few days, there were already 100 posts from everyone. Why? Because the last problem is so challenging, everyone turns to the forum to help each other. At first, I passed the advanced problem, but my algorithm running time was borderline. One fellow classmate suggests a brilliant improvement to my algorithm, and I was able to reduce it to 1% of the original running time! Crazy!
- The instructor is very active in the forum. He helped a lot of people with their problems, very frequently.
The drawback of this course is that it has pay-access (you will not be able to submit assignments without payment, however you can view them). However, I think it is very justifiable. It is very easy to tell that tons of effort, money, time and expertise have been put into making this course possible and it is well-worth the price tag.

I recommend this course, but it is not without its flaws. It has a lot of instructors and thus the quality of lectures varies a bit.

All of them do at least a passable job(all of them are top notch researchers by the way). The more engaging one, in my opinion, is Dr. Pavel Pevzner, he is very energetic and presents some interesting examples in the bioinformatics context, it's a shame that he teaches very few modules. The least engaging for me was also the one that presents the majority of the modules: Alexander S. Kulikov. The quality of his modules varies a lot. At times he seems like wasn't prepared to record the lecture and just confuses himself in the explanation. But overall, you can learn from it if you make the effort.

There were very few in-video quizzes and it is something that helps me keep on my toes while watching the lectures, so I think they could've added more. The assignment quizzes were not very great either, most of the questions do not present a challenge and overall didn't contribute much to my learning experience.

One thing that I like is that the slides are very well-crafted and often besides that, there are more advanced suggested readings.

I recommend the course to anyone that wants to learn about basic programming paradigms because, despite the shortcomings, the course has some spectaculars programming assignments. They resemble very much the questions from ACM-ICPC and judge your code based on correctness, and fitting in the proposed time and space constraints(a lot of allowed languages, but starter codes only for Python, C, C++ and Java). Often the more advanced questions even required some optimizations in the code to pass the bigger test cases within the time limits, so even if your solution is optimal, asymptotically speaking, it requires tweaking the code to get a passing grade.

I am unhappily surprised by the fact that some people rate this course low because it is allegedly poor explained. In my opinion, it is probably the clearest, best explained and best structured course among all algorithmic courses I've ever taken. Lecturers explain difficult things in the simplest terms. Even I was able to understand it, although I have very poor mathematical and programming background. Well, you will need to make some efforts to understand some concepts - but algorithms are hard to study just by default!

In addition, the course is highly adjustable for each student according to their level and/or desire to go deeper into details. it is easy enough to complete 1-3 challenges needed to pass, but you can also try to solve more difficult and interesting assignments.

I also like the way how all the statements and logic transitions are proved - every important point is proved in a very clear and easy to follow manner, you don't need to have a strong mathematical background to understand it. In other algorithmic courses, statements are usually proved either using complex mathematical terms or aren't proved at all.

I strongly recommend this course, it is just great!