IU Smart Course Selector AI Application

Inspiration

Choosing the right courses shouldn’t be a guessing game, especially when you’re studying abroad or navigating a new academic environment. Our journey as international students—often overwhelmed by an endless sea of course options—inspired us to develop a solution that cuts through the confusion. With personalized recommendations and insightful guidance, our platform helps you build an academic path that truly reflects your goals and dreams, empowering every IU learner to achieve success.

What it does

The IU Smart Course Selector AI Application is designed to:

Analyze a student’s academic history and interests.
Recommend courses tailored to individual profiles using advanced search techniques.
Provide detailed information about course prerequisites, descriptions, and student feedback.
Offer a conversational interface that allows students to explore courses, manage a personalized course list, and get real-time guidance.

How we built it

We combined several modern technologies and methodologies:

Backend: Developed using FastAPI integrated with a PostgreSQL database enhanced with the pgvector extension for semantic search.
Multi-Agent Framework: Implemented with LangChain, featuring specialized agents (Supervisor, Decision, Recommendation, Inquiry, Clarification) to handle and process user queries.
Frontend: Built with React, Redux, R and TypeScript to create a dynamic, interactive user interface.
Data Integration: Leveraged up-to-date course data from the IU Data Catalog Search.

Challenges we ran into

During development, we encountered several challenges:

Data Integration: Harmonizing diverse data sources and ensuring accurate, up-to-date information was a complex task.
Semantic Search Tuning: Implementing and fine-tuning the pgvector extension to deliver precise and relevant search results.
User Experience: Creating an intuitive interface that caters to the varied needs of students required multiple design iterations and continuous feedback.
Agent Coordination: Developing a seamless communication framework between multiple agents was challenging and necessitated extensive testing.

Accomplishments that we're proud of

Robust Multi-Agent System: Successfully built a framework that can intelligently process complex queries and deliver personalized responses.
Engaging UI: Developed an interactive, user-friendly interface that makes exploring courses and managing academic plans straightforward.
Enhanced Recommendations: Improved the relevance and accuracy of course suggestions through advanced semantic search.
Scalable Infrastructure: Created a backend that supports current needs while being adaptable for future growth.

What we learned

This project taught us valuable lessons:

Collaboration is Key: Integrating various technologies and coordinating across teams was essential for success.
User-Centric Design: Continuous user feedback helped refine the system to better meet student needs.
Optimization Balancing: Achieving the right balance between performance and precision in semantic search required significant fine-tuning.
Adaptability: Overcoming unforeseen challenges reinforced the importance of a flexible development approach.

What's next for Usecase - 2 : IU Smart Course Selector AI Application

Looking ahead, we plan to:

Enhance the Recommendation Engine: Integrate more advanced machine learning models for even better course suggestions.
Expand Data Sources: Incorporate additional academic metrics and student feedback to refine recommendations.
Upgrade the UI: Introduce more interactive features for a richer user experience.
Collaborate with Advisors: Explore partnerships with academic advisors to further personalize course recommendations.
Iterate Based on Feedback: Continuously gather and act on user feedback to improve the system.

Our goal is to evolve the IU Smart Course Selector into an indispensable tool that empowers students to make informed, personalized decisions about their academic futures.