CourseStack

AI plans out an engaging, personalized syllabus for the student.
Weekly lesson plans and modules.
Example of a problem set generated for the student, who has the ability to submit for a grade and check their work.
<3

Inspiration

The inspiration for this project came from recognizing the growing potential of AI as a powerful learning tool. We wanted to harness its versatility and ability to adapt to individual needs to create a more personalized educational experience. By leveraging AI, our goal was to provide learners with one-on-one guidance and tailored instruction, something that’s often difficult to achieve in traditional learning environments. At scale, this means giving every learner, regardless of background or resources, the kind of individualized support that was once reserved for the few.

What it does

CourseStack is an AI-powered learning platform that creates personalized courses based on a topic you want to learn. It asks targeted questions to understand your background, goals, and preferences, then builds and refines a custom curriculum for you to review and adjust. Once finalized, it generates a structured timeline with weekly plans, including extensive lecture notes, problem sets, and projects as well as interactive quizzes and exams. The user can interact with an AI professor at any time, and every curriculum design choice and all feedback is tailored to their preferences and learning.

How we built it

Frontend: We built the UI as a modern React app using Next.js (App Router), with server components and dynamic routes for weekly plans and lecture viewing. It handles course navigation, uploads, and displaying AI‑generated content with a clean, responsive design.

Backend: The backend is a Python service that orchestrates multiple AI workflows: it ingests course materials, runs specialized agents (for lectures, projects, and problem sets), calls external models and related utilities, and stores/structures results using the data models. The frontend talks to this backend through HTTP endpoints to trigger pipelines like lecture-note generation, graded assessments, and project scaffolding.

Challenges we ran into

We initially struggled in interfacing with the API, and had trouble getting the AI to stick to its role and produce output in structured formats. We also ran into limits in the AI's I/O ability, where longer lectures and assignments would fail to generate as they exceeded the max output token limit.

Accomplishments that we're proud of

Given its great variety of functionalities, we are proud of completing the full stack of CourseStack within the limited time, especially the automatically generated lecture notes and weekly plans.

What we learned

Full stack development is no joke, and don't be afraid to pivot from your initial ideas. Also, "Do NOT start the day off with 460mg of caffeine."

What's next for CourseStack

Eventually we would like to turn CourseStack into a free online service accessible to most people. We would also like to improve the multimodal capabilities of CourseStack and further improve the user experience interfacing with the lecture modules.