Qayyim: AI Learning Management System
THIS IS MY GRADUATION PROJECT!
Which I successfully defended and got a grade of A+ on January 28th 2026!
Overview
This was the most technically challenging project I have ever worked on and served as my graduation project, which I completed with a grade of A+. The App is a learning management system that leveraged LLMs to grade exams and provide personalised feedback to students. It consists of both a teacher UI and a student UI, complete with dashboards, exam review pages, a grievance management system, and a course enrollment system.
Project Description
AI-powered exam grading using OCR + LLM reasoning pipeline, RAG-based evaluation system using vector search (ChromaDB), parallel PDF processing using BullMQ queues, role-based authentication (JWT: student / instructor), course enrollment via secure join links, cloud-based deployment with Dockerized services, AWS S3 integration for exam storage.
My Contribution
Worked on most the backend and frontend of the website using Next.js. This includes, the database, integration of AI models and their endpoints into the frontend, the authentication system, the course-enrollment system, and the parallelization using BullMQ and Redus. Some parts of the website were done in collaboration with my colleagues such as the database, or by my colleagues such as the grievance system.
Reflection
The system had many complex components such as: OCR, JSON schema validation, vector search, and multi-tier grading logic, that were very tricky to synchronize across microservices, but it was a fun and technically enriching experience nonetheless!