📦 Product Workflow Overview

This section explains the end-to-end journey followed by a professor using Smart QnA — from logging in to sharing results with students. The system is optimized to minimize manual effort while ensuring high evaluation accuracy using AI.

🧭 High-Level Workflow

Below is a streamlined diagram of how a professor interacts with the platform:

📝 Step-by-Step Breakdown

1. 🔐 User Login & Role Assignment

• Professors, TAs, and Admins log in using their credentials.
• The system identifies their role:
  • IC (Instructor-in-Charge)
  • Co-Instructor
  • Teaching Assistant

2. 📚 Course Creation & Management

• Professors create a course by entering name, term, and department.
• Students are imported via CSV or added manually.
• TAs can be assigned to specific courses for assistance.

3. 🧪 Exam Setup

• Professors create exams within a course.
• Questions are uploaded with text/images, subject tags, and assigned marks.
• A golden answer script is uploaded as the reference solution.

4. 🧠 Rubric Generation

• AI analyzes the golden script and segments it into:
  • Written components
  • Diagrams
  • Tables
  • Equations
• Professors can annotate, adjust marks, and specify partial marking rules.

5. 📤 Upload Student Answers

• Professors upload scanned answer sheets in bulk (ZIP format).
• System auto-maps files to students using filename parsing.

6. 🤖 Auto Evaluation

• The Vision Transformer + LLM-based model compares student answers to the golden script.
• Scores are computed based on similarity and rubric configuration.
• If partial marking is enabled, proportionate scores are assigned.

7. 📬 Results & Feedback

• Professors review results on the grading dashboard.
• Students are notified automatically once grades are finalized.
• Detailed per-question feedback is accessible.

🔁 Role of Teaching Assistants

TAs can be granted access to:

• Assist in question uploads
• Help annotate golden scripts
• Review and validate AI-evaluated answers

✅ Outcome

With Smart QnA, educators save hours of manual grading while ensuring fairness and consistency.
The entire evaluation process is digitized, trackable, and highly scalable.