AI literacy test

Project type

Full-stack web application

Year

2025

Role

Full-stack developer (two-person team)

Team

2 — Saif Ali Rana and me

Technologies

MongoDB, Express, React, Node.js

Tools

Figma, GitHub, Postman

GoalBuild a platform, backend and frontend from scratch, that tests a user's ability to identify AI content and lets researchers configure tests and extract results.

AI-generated content is increasingly hard to distinguish from human-made content, and there was no simple way to measure how well people can actually tell the difference — or to collect data on it.

Target usersAnyone curious about their own AI literacy, plus researchers and educators who want to run and analyse tests.

• Reviewed the course requirements and mapped them to two distinct user groups: the participant taking the quiz and the researcher administering it.
• Compared stack options and chose MERN for a single-language codebase across client and server.

1. 01

   Requirements

   Broke the exam brief into participant-facing and admin-facing features.

2. 02

   Data model

   Designed schemas for quizzes, questions, media and submitted answers.

3. 03

   Backend

   Built the Express/MongoDB API with authentication for admin access.

4. 04

   Frontend

   Implemented the quiz flow and the admin panel in React.

5. 05

   Stats

   Added aggregation of submitted answers into exportable statistics.

6. 06

   Testing

   Ran the full flow end to end and fixed edge cases before delivery.

• Kept the participant experience as a stripped-back quiz so nothing distracts from the judgement being made.
• Separated the admin panel entirely from the public flow to keep configuration complexity away from participants.

• Built everything ourselves on the MERN stack — no scaffolding — including auth, quiz configuration and answer storage.
• Implemented aggregation queries so admins can pull statistics from all submitted answers.

• MongoDB
• Express
• React
• Node.js

A quiz platform where users judge whether content is AI-generated, backed by an admin panel where authorised users create quizzes and extract statistics from the collected responses.

Delivered a solution that met all requirements of the full-stack exam project.

What worked

Agreeing on the API contract early let both of us build client and server in parallel.

What did not work

We underestimated the scope and ended up with a lot of late nights close to the deadline.

What I learned

How to design and build a complete full-stack application, including auth and data aggregation.

What I would improve

Better planning up front, plus richer analytics and per-category scoring for participants.

My contribution

Built large parts of both the backend API and the React frontend, including the admin panel.

• GitHub repository ↗