midsummer madness - Tutorly AI

• 

  Architecture

Project Overview:
Our web application assists tutors in grading students' work and providing targeted feedback. Tutors upload their students' homework, which is marked and evaluated by AI, highlighting areas of improvement. This enables tutors to tailor their teaching and save time on repetitive grading, focusing instead on personalised teaching.

Key Functionalities: Our application features AI-assisted marking and automatic detection of student weaknesses from submitted work, allowing question generation of weak topics based on their needs. Our Tutor Dashboard allows tutors to quickly identify struggling students and manage each learner’s progress. Tutors can easily view submissions, work progress, and performance insights in one place.

Pain Points Addressed: Manual grading is time-consuming and prone to inconsistency by the tutor. Our solution automates evaluation with a standardised process, ensuring fairness and efficiency. By highlighting each student’s weak areas, the platform enables tutors to focus their teaching where it matters most, giving a more targeted and personalised approach.

External APIs: Our application leverages OpenAI for marking, while also using Supabase for authentication and database management.

Tech Stack Frontend: React, Vite, Tailwind CSS, Lucide-react Backend: Node.js, Express, Supabase JS SDK, OpenAI SDK, pdf-parse, mammoth, tesseract.js Database/Cloud: Supabase (Postgres, storage, auth) AI: OpenAI API (GPT-3.5/4)

Built With

• css3
• express.js
• lucide-react
• mammoth
• node.js
• openai
• pdf-parse
• react
• supabase
• tailwind
• tesseract.js
• vite