Pratap Tokekar | PlanetTerp

Information	Review
Pratap Tokekar CMSC421 Expecting a B+ Anonymous 06/05/2025	Lectures were pretty good, and the slides were very nice. He recorded lectures which was nice. Exams were fair and he allowed us to bring one page of notes for the exams. There were 4 written homeworks which were not released on a consistent basis, and one coding project at the end of the semester which was pretty easy. I wish the class was more organized and that there was 1-2 more coding projects. But overall a solid experience.
Pratap Tokekar CMSC421 Expecting an A- Anonymous 05/20/2025	Pratap is pretty good as a professor. He explains concepts in a coherent way, he's pretty understanding, and overall just a solid professor. Any complaints I have are directly against the organization/logistics of the class. Practice exams had some wrong answers which tripped me up, and responses on Piazza vary a LOT. Usually, it's on the much longer end. Class content itself is alright. I did learn a lot, but the content is directly taken from Berkeley's class. It's funny too; I looked up something from the slideshow for more information about that topic and all I found was the same slideshow on 4 different universities' websites. I'd recommend taking Pratap, and probably 421. Class is relatively easy too.
Pratap Tokekar CMSC421 Expecting an A Anonymous 05/18/2025	By far the best option you have for CMSC421 (though imperfect) Pratap is quite a nice lecturer. When looking at the slides at first (which, as AI is, is still stolen from Berkeley), some equations and concepts are kind of hard to grasp at first glance. But Pratap explains it in a simplified way while not removing the details fully. Although some of the AI concepts at the start are repetitive (BFS, DFS, and other stuff you learned), there's just enough to make the start somewhat different. Pratap records all his lectures (which lowers attendance drastically, but near the end of the semester, some lectures stopped being recorded for some reason. There were intended to be more assignments, but likely to disorganization, there were only 4 homework assignments and 1 lab. I really was hoping to do more coding (and the coding assignment was pretty easy, though these 5 assignments ended up being weighted up 10% each, crazy given the homework was also short). This may have happened due to disorganization, and really, disorganization is the main gripe I have with this class. TAs don't answer on Piazza, and Pratap answers nearly every question (I feel bad for him qwq). Homework grades come quite late, regrades are annoying, and questions are written strangely quite often. I feel that if this class were managed much better, it would be perfect, and strangely enough, more homework. Tests are easy. Like, really easy. The practice exam is extremely similar to the actual exam, so similar that it incentivizes overfitting on the practice exam (and, plus, you even have a cheatsheet for each test!). The averages were extremely high (87ish?) for each test, so that will not be a concern unless something changes. Overall, given the teachers I've seen teaching AI, be glad you have Pratap (though he's going on sabbatical...). Despite some flaws, he runs a good course for a notoriously poorly run class.
Pratap Tokekar CMSC421 Expecting an A Anonymous 05/14/2025	loved taking cmsc421 with Pratap. him teaching it definitely made it the favorite upper level cs class that i've taken at umd. slide decks are structured very well and provide a ton of information, all of which i found really interesting. homeworks and projects are very interesting to work on and he provides practice exams for his exams, so you know what to expect on exams and can prepare really well for them. 10/10 experience for me and would highly recommend taking cmsc421 with Pratap to anybody.
Pratap Tokekar CMSC818B Expecting an A+ Anonymous 12/04/2024	I loved taking this class with Pratap. He explained the concepts very well, there wasn't too much work, and the grading was lenient. However, there could have been better organization.
Pratap Tokekar CMSC421 Expecting an A- Anonymous 12/22/2022	I made a review back in November regarding Prof Tokekar and I would say that his management of the 421 course improved between then and the end of the semester. Pros: 1) Posts lecture slides and links to additional online resources and textbook chapters to review 2) Mix of online and in-person OH and TAs did a good job at answering questions during OH 3) Interesting assignments and projects 4) Answers questions during lecture, OH, and is very flexible 5) Posts practice exams and problems to prepare for final and midterm 6) Allows regrade requests 7) Quickly graded final exams Cons: 1) Took a while for midterms to be graded and transferred to ELMS 2) TAs took a little while to respond to Piazza posts
Pratap Tokekar CMSC421 Expecting an A Anonymous 12/10/2022	I think the professor taught as best as he could, but the class structure just wasn't good at all. The TAs never responded on Piazza so he was basically the only one who ever answered questions, grades took forever to get back, lectures weren't recorded for unexplained reasons, and the slides were taken from UC Berkeley and never updated. Pratap has potential to be a great professor, but if he's gonna teach this class in the future he should just make his own slides, and he should push the TAs to at least answer student questions.
Pratap Tokekar CMSC421 Expecting a B Anonymous 11/10/2022	Professor Tokekar is a well-intentioned professor with a strong background in AI and the material within the course is interesting. He also does a good job at creating assignments and projects that are relevant to the concepts we were studying. He also answers questions well within lecture and breaks down the topics for exams and best ways to approach projects/assignments. The TAs also do a good job at answering questions during online & in-person office hours. However, there are a few things that I think Prof Tokekar could improve upon when it comes to this course: 1) Posting the lecture slides in a more timely manner because sometimes, it takes more than 1 day for the most recent lecture slides to be posted and students that missed lecture may have to wait to catch up 2) Responding to Piazza posts more quickly as it takes a while for students to get answer to those questions 3) Posting more practice problems & exams as although the assignments do a good job at covering the concepts and the lecture slides have good examples, more problems would help us master these concepts easier. Also, we only got 1 sample midterm and having at least 1 more would have helped. 4) Faster grading & regrading because it took almost 4 weeks for midterm results to be released onto ELMS and a few weeks for projects to be graded and is taking a while for grades to be transferred from Gradescope to ELMS Other than that, I would say the course was managed well by Professor Tokekar.

Pratap Tokekar
CMSC421

Expecting a B+
Anonymous
06/05/2025

Lectures were pretty good, and the slides were very nice. He recorded lectures which was nice. Exams were fair and he allowed us to bring one page of notes for the exams. There were 4 written homeworks which were not released on a consistent basis, and one coding project at the end of the semester which was pretty easy. I wish the class was more organized and that there was 1-2 more coding projects. But overall a solid experience.

Pratap Tokekar
CMSC421

Expecting an A-
Anonymous
05/20/2025

Pratap is pretty good as a professor. He explains concepts in a coherent way, he's pretty understanding, and overall just a solid professor. Any complaints I have are directly against the organization/logistics of the class. Practice exams had some wrong answers which tripped me up, and responses on Piazza vary a LOT. Usually, it's on the much longer end. Class content itself is alright. I did learn a lot, but the content is directly taken from Berkeley's class. It's funny too; I looked up something from the slideshow for more information about that topic and all I found was the same slideshow on 4 different universities' websites. I'd recommend taking Pratap, and probably 421. Class is relatively easy too.

Pratap Tokekar
CMSC421

Expecting an A
Anonymous
05/18/2025

By far the best option you have for CMSC421 (though imperfect) Pratap is quite a nice lecturer. When looking at the slides at first (which, as AI is, is still stolen from Berkeley), some equations and concepts are kind of hard to grasp at first glance. But Pratap explains it in a simplified way while not removing the details fully. Although some of the AI concepts at the start are repetitive (BFS, DFS, and other stuff you learned), there's just enough to make the start somewhat different. Pratap records all his lectures (which lowers attendance drastically, but near the end of the semester, some lectures stopped being recorded for some reason. There were intended to be more assignments, but likely to disorganization, there were only 4 homework assignments and 1 lab. I really was hoping to do more coding (and the coding assignment was pretty easy, though these 5 assignments ended up being weighted up 10% each, crazy given the homework was also short). This may have happened due to disorganization, and really, disorganization is the main gripe I have with this class. TAs don't answer on Piazza, and Pratap answers nearly every question (I feel bad for him qwq). Homework grades come quite late, regrades are annoying, and questions are written strangely quite often. I feel that if this class were managed much better, it would be perfect, and strangely enough, more homework. Tests are easy. Like, really easy. The practice exam is extremely similar to the actual exam, so similar that it incentivizes overfitting on the practice exam (and, plus, you even have a cheatsheet for each test!). The averages were extremely high (87ish?) for each test, so that will not be a concern unless something changes. Overall, given the teachers I've seen teaching AI, be glad you have Pratap (though he's going on sabbatical...). Despite some flaws, he runs a good course for a notoriously poorly run class.

Pratap Tokekar
CMSC421

Expecting an A
Anonymous
05/14/2025

loved taking cmsc421 with Pratap. him teaching it definitely made it the favorite upper level cs class that i've taken at umd. slide decks are structured very well and provide a ton of information, all of which i found really interesting. homeworks and projects are very interesting to work on and he provides practice exams for his exams, so you know what to expect on exams and can prepare really well for them. 10/10 experience for me and would highly recommend taking cmsc421 with Pratap to anybody.

Pratap Tokekar
CMSC421

Expecting an A-
Anonymous
12/22/2022

I made a review back in November regarding Prof Tokekar and I would say that his management of the 421 course improved between then and the end of the semester. Pros: 1) Posts lecture slides and links to additional online resources and textbook chapters to review 2) Mix of online and in-person OH and TAs did a good job at answering questions during OH 3) Interesting assignments and projects 4) Answers questions during lecture, OH, and is very flexible 5) Posts practice exams and problems to prepare for final and midterm 6) Allows regrade requests 7) Quickly graded final exams Cons: 1) Took a while for midterms to be graded and transferred to ELMS 2) TAs took a little while to respond to Piazza posts

Pratap Tokekar
CMSC421

Expecting an A
Anonymous
12/10/2022

I think the professor taught as best as he could, but the class structure just wasn't good at all. The TAs never responded on Piazza so he was basically the only one who ever answered questions, grades took forever to get back, lectures weren't recorded for unexplained reasons, and the slides were taken from UC Berkeley and never updated. Pratap has potential to be a great professor, but if he's gonna teach this class in the future he should just make his own slides, and he should push the TAs to at least answer student questions.

Pratap Tokekar
CMSC421

Expecting a B
Anonymous
11/10/2022

Professor Tokekar is a well-intentioned professor with a strong background in AI and the material within the course is interesting. He also does a good job at creating assignments and projects that are relevant to the concepts we were studying. He also answers questions well within lecture and breaks down the topics for exams and best ways to approach projects/assignments. The TAs also do a good job at answering questions during online & in-person office hours. However, there are a few things that I think Prof Tokekar could improve upon when it comes to this course: 1) Posting the lecture slides in a more timely manner because sometimes, it takes more than 1 day for the most recent lecture slides to be posted and students that missed lecture may have to wait to catch up 2) Responding to Piazza posts more quickly as it takes a while for students to get answer to those questions 3) Posting more practice problems & exams as although the assignments do a good job at covering the concepts and the lecture slides have good examples, more problems would help us master these concepts easier. Also, we only got 1 sample midterm and having at least 1 more would have helped. 4) Faster grading & regrading because it took almost 4 weeks for midterm results to be released onto ELMS and a few weeks for projects to be graded and is taking a while for grades to be transferred from Gradescope to ELMS Other than that, I would say the course was managed well by Professor Tokekar.