Inspiration

I am a data scientist and tutor, and education has always been personal to me. I came through the O Level and A Level system myself, so I understand the pressure of final exams, completing past papers, and waiting to know whether you are actually ready.

As a tutor, I also saw how much time is lost marking papers manually. Students need feedback quickly, but tutors and teachers have limited time. Many existing platforms are expensive, generic, or only provide practice questions without showing students where they truly stand.

That inspired ExamPulse: an accessible exam-readiness platform that turns past-paper practice into a clear, motivating readiness journey.

What it does

ExamPulse helps O Level and A Level students:

• Select their subject, syllabus, paper year, series, and paper number
• Practice using past-paper-style assessments
• Submit answer sheets for AI-assisted evaluation
• Receive subject-specific feedback and improvement guidance
• Track readiness, identify weak topics, and build momentum
• Join a subject league to make preparation feel motivating rather than isolating

The aim is not just to give a mark. It is to help students understand:

“Am I ready for my exam, what should I improve, and what should I do next?”

How we built it

I built ExamPulse as a modern web application using React, Next.js/Vinext, TypeScript, and Cloudflare-ready services for storage and records.

The platform includes:

• A subject-aware dashboard for Mathematics and Physics pathways
• A paper-bank and custom-paper assessment flow
• AI-assisted answer-sheet analysis using GPT-5.6
• Readiness insights, topic feedback, and improvement actions
• Student identity, submission history, and league features
• A local preview mode so the project can be demonstrated without requiring a complex setup

Codex was central to the build process. I used it to accelerate UI development, structure the assessment flow, refine the student experience, create the API routes and data model, debug local authentication issues, write tests, and produce the documentation needed to run the project.

Challenges we faced

The biggest challenge was making the platform feel like a useful product rather than a simple AI demo. A student should not need to understand AI to benefit from it, so the experience had to be simple: choose a paper, submit work, and receive an actionable readiness view.

Another challenge was supporting multiple subjects without creating separate versions of the product. We solved this by making the dashboard, proof runs, paper selection, and feedback subject-aware.

We also had to make local development reliable. Hosted authentication and cloud storage are not naturally available on localhost, so we added a safe local-preview mode that lets judges explore the experience while preserving the production architecture for real AI marking and student records.

What we learned

We learned that the most valuable part of AI in education is not just generating answers. It is turning evidence of a student’s work into timely, understandable next steps.

We also learned that exam preparation is emotional as well as academic. Students often know they need to study, but they do not know what to prioritize. ExamPulse makes readiness visible, breaks improvement into manageable actions, and adds a light competitive layer to make consistent practice more motivating.

What’s next

My long-term goal is to launch ExamPulse as a cost-effective platform for students and tutoring businesses. It can reduce the time tutors spend marking papers while giving students faster, more personalised feedback.

I want to expand the paper bank, support more curricula and subjects, improve handwritten-answer evaluation, and give tutors a view of class-wide readiness so no student enters an exam without knowing where they stand.

Built With

• ai
• cloudflare
• codex
• edtech
• education
• exam-preparation
• gpt-5.6
• handwriting-recognition
• nextjs
• react
• student-productivity
• typescript