Exam Paper Generator

• 

  Front Page

• 

  Practice Exam Page

• 

  Generation Parameter Page

• 

  Guided Interactive Practice Experience

• 

  Front Page

Inspiration

Students often find themselves wishing for practice exam papers for their university courses. However, not all courses provide them. That’s why we envisioned a tool powered by advanced AI that allows every student to generate high-quality practice exams based on their course slides. Our goal is to make exam preparation more accessible, efficient, and tailored to individual learning needs.

What it does

The tool enables users to upload their lecture slides and instantly generates an interactive practice exam with a single click. The generated content is high-quality and custom-tailored to help students effectively review and practice material directly from their slides. By transforming static lecture content into dynamic, engaging assessments, we empower students to reinforce their understanding and improve their exam readiness.

How we built it

We developed a custom, advanced data pipeline designed to extract content from lecture slides and convert it into structured information. This data is then processed by a sophisticated large language model (LLM) workflow, which generates a set of practice questions based on predefined parameters.

To ensure the highest quality and accuracy, we implemented a multi-agent system that optimizes question generation and ensures the content closely matches the input materials. This system refines and validates the generated questions, making sure they align with the key topics covered in the lecture slides.

Challenges we ran into

Building an AI-driven product requires seamless collaboration across multiple domains, which presented a significant challenge for our team.

Our AI research team focuses on model development, fine-tuning, and optimization. The backend team ensures efficient system integration and data processing. The frontend team designs an intuitive user interface for a seamless user experience. Each team operates with different priorities, workflows, and technical expertise. For example, AI researchers work iteratively on model improvements, which may require computationally expensive processes, while the backend team needs to integrate these models into a scalable system. Meanwhile, the frontend team relies on structured APIs and well-organized data to build a user-friendly interface.

Bridging these differences required clear communication, well-defined workflows, and collaborative tools to keep everyone aligned on goals, expectations, and progress.

Accomplishments that we're proud of

A Custom AI Pipeline for Practice Question Generation Capable of processing large lecture files containing hundreds of slides Generates high-quality, well-structured practice questions

An Intuitive and Visually Appealing User Interface Designed for ease of use and smooth interaction Provides an engaging and interactive exam experience

What we learned

• Collranoration is the key of success
• Understand the need of user is important
• There is a lot of things to do to optimize the quality of the LLM output

What's next for Exam Paper Generator

• Provide planning feature to help students create study plan to practice thier units
• Provide insightful advice on which sections of contents the students need to revise

Built With

• langchain
• langgraph
• langsmith
• nextjs
• react