Qnectify
An AI-powered EdTech platform that automatically generates quizzes from PDF study materials. Built with a decoupled architecture, it uses generative AI to save teachers time and gamifies the student experience to increase engagement.
The Challenge
Teachers spend hours manually creating quizzes from textbooks. Existing auto-generators often produce irrelevant questions or hallucinate facts, leading to poor learning outcomes for students.
The Solution
We developed a system where the Frontend (Next.js) and Backend are separated to ensure scalability. I integrated the Gemini API with a robust prompt engineering layer to extract accurate questions from uploaded PDFs, while the backend manages the gamification logic (XP, Leaderboards) securely.
Key Features
Technology Stack
Next.js
Frontend UI and Client-side interactions.
Node.js
Backend API logic and Database management.
Gemini API
Generative AI for content creation.
Tailwind
Responsive UI styling.
Development Process
Research & Analysis
Analyzed the pain points of manual quiz creation and gamification strategies.
Design System
Created high-fidelity designs focusing on an engaging, game-like user interface.
Parallel Development
I worked on Slicing the UI components while simultaneously developing the Backend API endpoints.
API Integration
Fetching data from the backend and connecting the Gemini AI service.
Testing
Validating AI response accuracy and fixing bugs in the scoring logic.
Finishing & Presentation
Final polish and pitching the project at HOLOGY 8.0.
Engineering Challenges
AI Hallucinations
The AI sometimes generated wrong answers. I added a validation step where the AI must cite the page number from the PDF.
State Management
Syncing the gamified state (score/streak) between the decoupled frontend and backend required robust API error handling.