|
Description: |
Project Overview: We are looking to develop a web-based application that allows an admin to upload PDFs on various topics and generate quizzes based on the uploaded content. The quizzes should support multiple question types and leverage generative AI (GenAI) for question generation and free-text answer evaluation. The platform should be responsive for mobile usage, with a future possibility of developing a dedicated mobile application.
User Roles:
1. Admin
o Upload PDFs categorized by topic.
o Generate quizzes using GenAI.
o Review and edit quiz questions before publishing.
o Configure quiz settings.
o View reports.
2. User
o Attempt quizzes.
o Receive feedback on free-text answers.
o View quiz results.
Quiz Functionality (Minimum Requirement):
• Questions generated using GenAI from uploaded PDFs.
• Supports different question types:
o True/False
o Multiple Choice
o Fill in the Blanks
o Free Text
• Admins can manually add/edit questions before publishing.
• For free-text responses, the LLM generates an "ideal" answer and evaluates the user’s response based on key points, providing a score and feedback.
• A minimum threshold (e.g., 70%) can be set to mark free-text answers as correct.
Note: The above represents the minimum requirement. If additional question types can be easily implemented, we are open to considering them.
________________________________________
Admin Configurable Settings:
• Number of questions to generate.
• Number of questions presented to the user per quiz.
• Pass percentage.
• Timed Exam (Yes/No).
• Randomization of questions and response options.
________________________________________
Technologies & Infrastructure:
• Backend: Java (preferred)
• Frontend: Open to developer’s recommendation
• LLM Options: OpenAI, Perplexity, Grok, Gemini (whichever is best suited)
• File Storage: AWS S3 (for uploaded documents)
• Hosting: AWS (deployment support required)
________________________________________
Reporting Requirements:
• User quiz performance tracking.
• Breakdown of correct/incorrect answers.
• Analysis of free-text response scoring.
Expected Deliverables:
1. Fully functional web-based quiz application.
2. Admin interface for quiz generation, review, and configuration.
3. User interface for quiz participation.
4. AI-powered question generation and free-text evaluation.
5. Reporting dashboard for quiz analytics.
6. AWS deployment and setup support.
Project Expectations from Developer:
• Provide a clear approach to implementing AI-based quiz generation and evaluation.
• Ensure a user-friendly, responsive web interface.
• Outline the estimated timeline for delivery.
• Detail technology choices and rationale.
• Ensure scalability and security best practices.
Note: Generic responses will not be entertained. The proposal should demonstrate an understanding of the requirements and a structured approach to execution. |
