Overview

An AI-powered adaptive learning system developed as a hackathon project. VideoQuiz AI transforms YouTube educational videos into topic-based quizzes that adjust in difficulty as learners answer questions, then analyzes their performance to identify strengths and areas for improvement.

The challenge

Watching an educational video does not necessarily mean that the material has been retained or understood. Learners may struggle to identify gaps in their knowledge, while manually creating and grading meaningful revision questions takes time.

Approach

The pipeline began with a YouTube video URL and retrieved its transcript. The text was cleaned to remove irrelevant material such as introductions, subscription requests, and promotional segments. The system then identified important content, organized questions by topic, and generated questions using a separate question-generation workflow. An adaptive grading mechanism increased the difficulty as learners progressed and used their responses to analyze different areas of understanding.

Key features

• YouTube URL-based learning workflow
• Video transcript retrieval
• Transcript cleaning and removal of irrelevant segments
• Identification of important concepts
• Topic-based question generation
• Difficulty progression based on learner responses
• Adaptive question grading
• Analysis of stronger and weaker subject areas
• Insights into understanding, analytical ability, and recall

Takeaway

VideoQuiz AI explored how generative AI can turn passive video consumption into a more interactive learning experience. The most important part was not simply generating questions, but connecting content extraction, topic organization, adaptive difficulty, grading, and performance analysis into one learning pipeline.