AI Agents

I think a heavy emphasis on paper presentations dragged on a bit and i wish we had more exercises/shorter assignments and more time for final projects

I think some of the initial lectures could have been a bit more comprehensive. A lot of students come into this class with varying backgrounds and familiarity with the tools. I was able to follow along quite well, but some of the more technical readings and parts of the assignments had me spending a lot of time searching for what I needed to know. Totally fair as a more research oriented class, but I think having that a bit more streamlined during a 10 week term would be super helpful.

Nothing.

more teaching and a few less poresentations but loved it

Future iterations may benefit from refinement of the overall process, but ultimately this course would need to be a dynamic one.

It would be helpful to have more reminders for the reading assignments so students do not accidentally miss them. In addition, more instructor-led lecture content during class could further enhance the course, as currently a large portion of class time is devoted to student presentations.

The grading. There's limited understanding on the rubric for grading.

The workload is a bit too heavy

Towards the end of the term the course felt very overloaded by the student paper presentations, which we read for homework anyways. I would have liked to see more instructor-led exploration of material past the readings.

Unfortunately, the onslaught of student presentations in the later half of the course were not the most conducive to learning the concepts.

spread out pset1 into 2 assignments, it was a lot of different topics that could have been split into 2 problem sets

- Super interesting class

I think focusing on what agents fundamentally are, divorced from their implementation was important.

The homeworks were a lot of fun and very applied to the real world. I honestly feel like I learnt the material a lot better when I tried to apply it to real world problems.

The presentation, which included discussion-based elements, was among the most effective aspects of the course, as it allowed students to think deeply about the strengths and limitations of a paper, reflect on those limitations and tie them to other papers, and learn how to read papers effectively.

The assignments had an excellent balance of theory with practical implementations (and the right amount of structure but room to experiment). The research backing was well curated and explored.

The assignments were well designed and provided valuable hands-on experience in building AI agent systems. The paper presentation sessions were also very effective, allowing us to learn about recent research on AI agents and benefit from discussions with classmates.

The early lectures were especially helpful in establishing the foundation for AI agents. The readings were also useful in this regard.

The practical experience gathered throughout the class.

The problem sets were great opportunities to get my hands dirty and work with real-world implementations of agents.

Up to date knowledge on AI agent

paper presentations helped to reinforce what we learned through the readings

A lot of reading. The class readings were a lot at times, but extremely informative and comprised of a great foundational reading list. Most of the applied stuff was easily figured out with AI assistance when I got blocked.

I did all the assignments.

reading papers outside of class

By actively conducting research on course topics

I completed the assignments seriously, attended over 80% of the classes, and actively asked questions during discussions. This helped me stay engaged with the course and deepen my understanding of the material.

I put a strong effort towards the assignments and project.

Looking up related references in the papers that interested me in the required readings.

Through challenging assignments

Working on the assignments and projects, reading, and engaging in class discussions.

I thought the course was paced quite fast. There were a lot of readings and material to ramp up on in a short amount of time. That being said, it was a very fun and interesting course that made me want to keep up with the pace. I ended up learning a ton.

It was paced well.

hard class but yes it was good

A bit fast

Exceptional

I found the pace of the course reasonable. By reviewing the syllabus and following the instructor’s schedule, I was able to understand the overall structure of the course and keep up with the learning pace.

Moderate

Quite manageable but may be difficult for undergrads.

The pace of the course was good.

The pace was perfectly fine.

well paced, but please try to limit the amount of time ppl took during paper presentations, some realllyyy dragged on

I honestly didn't use the TAs too much as a resource as I found my own reading and reflection to be more helpful to digest the course material. When I had interesting questions or thoughts, the professors office hours were more than enough.

Not very helpful

Very helpful!

Exceptional

Prof. Nikhil taught us a lot about trending of AI

TAs were a bit late to grade assignments and some scores in gradescope and canvas dont match properly; could be improved. Alot didn't have alot of understanding for the psets.

The course staff, including the TAs, were helpful overall. One area that could be improved is the timeliness of grading. Earlier feedback from the TA team would allow students to better understand their performance and learn more effectively from the assignments.

They were great