Dialogue-Based Student Portal Design using a Large Language Model (LLM)

Supervisor

Dr. Nafis Tanveer Islam
📧 [email protected]

Description

A school wants to implement a dialogue-based “Student Management Portal”, allowing students to chat with a system regarding their course results, graduation requirements, and areas for improvement using microservices.

The portal should:

• Allow students to ask about their results.
• Help them identify areas for improvement.
• Show the number of remaining courses required for graduation.
• Ensure the system does not get overwhelmed due to excessive queries, leading to high costs.

---

Requirements

Core Functionalities:

1. Develop a secure student portal allowing students to log in.
2. Maintain individual student records in any chosen format (Relational DB, Non-Relational DB, CSV, JSON, etc.).
3. Ensure students cannot access another student’s performance data.
4. Provide a User Interface (UI) for profile viewing and editing.
5. Implement a Chat interface where students can:
   • Ask about their performance.
   • Check their scores in different subjects.
   • View remaining courses required for graduation.
   • (⚠️ Ensure cost estimation and monitoring of LLM usage).

---

Expected Output

Minimum Requirements:

✅ A working cloud application with:

• A secure, scalable, and user-friendly student portal.
• Integration of a dialogue system using LLM (e.g., OpenAI, Hugging Face).
• Login & profile management with information exclusivity for students.
• A Monitoring System to track usage and restrict excessive queries.

Advanced Deliverables (If Time Allows):

1. Identify and analyze performance bottlenecks:

   • Does the system slow down with multiple concurrent users?
   • Explain where and why the delay occurs.

2. Optimize and speed up the dialogue system.

3. Implement a cost-control protocol to limit excessive traffic.

---

References

1. Student Result Management System
2. Falcon-LLM Education Platform

This application is based on an example project from streamlit