professor
Sambit Sahu

May 2021

One of the most disorganized courses I've taken at Columbia! The professor provides an entire plan for the semester in the first lecture, but you can expect him to not follow it at all. Few examples of this: - He gave a distribution of the grades, and then during the lecture, he felt that he needs to add some more difficulty to the course. So he decided to add few research paper readings every week, and said it will carry 5% of the grade. But he never released the new split, and after a few weeks, he got bored and stopped giving questions on those research papers (Eventually none of those answers were corrected and those submissions carried 0 weight in the grade). - For the first assignment, he added an extra credit task. For the following assignments, nothing was released. When the grades for the 2nd assignment were released, one team had gotten 105/100 points, even though there was no extra credit component. On asking the TAs, they told us that one team had approached the Prof separately and asked him for an additional task, so he gave only them the extra credit. Other people in the course were never informed about this task, nor given any opportunity to get any such extra credits. After a lot of complaints by the students, he released the extra credit portion of the assignment for all the students and asked them to submit it 1 week after the finals. This is certainly not practical because students are busy with the final project work at that time, and everyone deletes all resources used for an assignment after the demo to avoid being charged by AWS. - For the project demo, there was no proper schedule followed, and he kept running late. Eventually, the last few teams for the day had to come back again the next day and present, having no regard for the students schedules. The lectures too are dull, with Professor Sahu being a really poor lecturer, didn't attend a single lecture after one point of time. The guest lectures in the course are good. Moving away from the poorly managed course, which are due to flaws of the Professor, there is a lot to learn for people who have never used AWS services before. If you have experience working in AWS, then this would be useless, as it doesn't really teach Cloud Computing or Big Data concepts, rather just how to implement projects on AWS. So I would still recommend this to students who have never used AWS before, as it would definitely be useful in the future. The grading is also lenient, and it is easy to get an A. But don't expect any sort of transparency from the Professor, and be prepared to get really annoyed by his behavior.

May 2021

Sahu sells this class as a mixture of hands-on, practical experience using Amazon Web Services, as well as the general theory behind cloud computing and distributed systems. During the semester, we were assigned 3 pretty meaty assignments involving AWS and its associated services. You can work with a partner on these, which I would recommend, but still expect to get bogged down in weird little bugs due to the alien set up of AWS and lackluster documentation. But these projects are pretty cool on their own. You get to build a recommendation bot, an image storage site, and a machine learning app, so there's a wide exposure to the different services that AWS offers. On top of this, there is a final project due around finals time, which was definitely stressful. There's also 2 pretty reasonable but time-consuming quizzes and weekly readings on abstruse technical topics that are borderline indecipherable to a newbie (mapreduce, google file system…). While this class is a lot of work, no one's trying to trick you, and you'll probably get a good grade as long as you don't slack off. As for the content, this class could open up a lot of opportunities for anyone interested in web development/cloud. AWS is an in-demand skill, and Sahu is not shy about sharing his experiences of students telling him how they got internship offers because of his class. The theoretical stuff was hard to understand but still worthwhile, and I definitely want to experiment with Hadoop/spark and Kafka more in the future. My major gripe with the course is that it was somewhat disorganized, probably due to the shortened, online semester because of COVID. The deadlines for the assignments and project often got pushed around. The lectures themselves are sometimes unclear and meandering, and I was often left wondering how any of it was relevant to our homework. There wasn't as much direction for the final project as I would have wanted, especially since there were necessary topics (frontend development and authentication) that weren't covered extensively during recitations. There are also no credits for this class, so you'll probably have to spend some cash out of pocket for your AWS account. Overall though, if you have time for it and want to get experience in the cloud, I would definitely recommend this course.

Jan 2017

The class is all about using AWS services to make web apps. You just need to make things work and show them to get marks. Lots of time is wasted in debugging and doing silly things that don't teach you cloud computing, but just DEBUGGING. The cloud concepts are somewhat covered in lectures partially and you are expected to learn yourself too through readings assigned every week. Overall, a good but unorganized course with more workload than it is supposed to have. You will definitely learn about many cloud technologies and will be able to make simple cloud apps that can impress someone. If you have time to code and do a super project oriented course in a semester, take this course!

Dec 2014

The class was basically a walkthrough of AWS. By the end you will have seen a bunch of different AWS services and this is kind of interesting but it should be approached as a teach-yourself-as-you-go type of class. Professor Sahu is an ineffective lecturer. He does provide a list of interesting papers to read, and recommends several other readings which are good, but I took away almost nothing from his lectures. I would recommend never showing up on time because nothing ever happens in the first 20 or so minutes. Additionally he missed several lectures, and had other guest lectures where clearly he had dropped it on the person at the last minute. I got the impression he had a lot of other things going on and this class was not his priority.

Nov 2014

If you want to learn anything avoid this professor at all costs! He is completely disorganized, does not respond emails and does not show up for office hours. He gets to class completely unprepared, then he spends around 10 minutes writing in his laptop and when he finally starts the class he shows the students he just came up with some assignment (that he wrote fairly bad in a Google Docs document). Then he spends 40 minutes trying to explain what he wants us to do, but nobody understands because he did not plan the assignment and frequently end up changing the assignment while he tries to explain. When it is finally time to start class for real he just starts talking about a topic, but often he gets confused and then decides to open the Wikipedia page of that topic to see if he can solve his confusion. He nevers go deeper than the Wikipedia page of a given topic, then spends the final minutes of the class basically just listing books and websites we should read for next week. That's right BOOKS. He once suggested we read 3 300 pages long books for the next week. Of course nobody reads a single page of that books and the professor doesn't even remember he suggested us to read the books afterwards. Anyway it is a really depressing course. I registered thinking I would learn something about building scalable systems and Hadoop and stuff. The most advanced thing we learned was the first most basic example of doing a wordcount with Hadoop. That was basically the last assignment as well. And the final project? Well it is something you will spend dozens of hours but you will only use what you already knew before. Basically building some web app that is hosted on Amazon and uses DynamoDb.