Loading...

Media is loading
 

Degree

Bachelor of Science (Computer Science)

Department

Department of Computer Science

School

School of Mathematics and Computer Science (SMCS)

Advisor

Ms. Abeera Tariq, Lecturer, Department of Computer Science

Keywords

Agentic AI, Automated Assessment, Educational Technology

Abstract

This project addresses the burden of repetitive administrative work faced by university educators on Learning Management Systems (LMS) by delivering two integrated AI-powered systems. The first is a browser automation agent that accepts natural-language instructions and uses Playwright with multi-LLM support (OpenAI, Anthropic Claude, Google Gemini) to autonomously navigate and perform tasks on browsers, such as uploading materials and managing courses, eliminating manual interaction. The second is an automated paper marking system that uses AI vision to evaluate scanned exam scripts against uploaded rubrics, generate per-question marks, and export results, reducing hours of manual grading to minutes. Together, the systems form a teacher productivity platform that demonstrates practical LLM-driven automation applied to real educational workflows, backed by a shared Supabase database for authentication and persistent data

Tools and Technologies Used

Python, FastAPI, Uvicorn, Playwright, OpenAI API, Anthropic Claude API, Google Gemini API, React 18, Next.js 14, TypeScript, Tailwind CSS, Supabase (PostgreSQL), WebSockets, Pydantic, PyPDF2, pdf to-img, Radix UI, Vite

Methodology

The project followed an iterative, feature-driven development approach. Work was broken into phases: (1) core Playwright browser automation engine with LLM tool calling integration; (2) FastAPI backend refactored into modular routers for browser control, auth, LMS data, and flow history; (3) React frontend with real-time WebSocket event streaming; (4) Supabase auth and LMS onboarding wizard so teacher credentials auto-inject into every agent task; (5) a separate Next.js paper marking application using AI vision to process PDF exam scripts and Jupyter Notebooks against marking rubrics. Both systems were developed in parallel and share the same Supabase project for auth and storage.

Document Type

Restricted Access

Submission Type

BSCS Final Year Project

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS