Created by: Jumbos: Into the Hackerverse (Katie Yang, Spencer Anderson, Hanah Kim, Emily Li, Madeline Lei, Tomas Maranga) Presented at JumboHack 2025 under the Down the Rabbit Hole track.
Our project is a two-fold solution designed to modernize academic advising. On one side, we've developed an advising chatbot interface that both students and advisors can use to get quick, informed responses to academic policy questions. On the other, we've built an Outlook plugin that leverages our API and retrieval-augmented generation (RAG) system to automatically generate draft responses from advisor emails, saving time and ensuring consistency in communication.
Team Member Contributions:
Katie Yang: Spencer Anderson: Worked on retrieveing and sending emails. Hanah Kim: Chatbot interface. Emily Li: Lead Designer, created presentation. Madeline Lei: Working on generating responses for the chatbot Tomas Maranga: Working on integrating APIs, making our own api routes.
ACKNOWLEDGEMENTS React & Flask: Robust frameworks, made our frontend/backend development smooth Microsoft Graph: Providing seamless email integration capabilities RAG: Used existing RAG model
REFLECTION From lengthy, confusing degree audits to lists of major requirements, making sure you are on the right path to graduate can be confusing. This leads to an influx of repetitive emails for academic advisors and distress for students. We wanted to create a solution to help students and advisors navigate their academic journey here at Tufts.
In the future, we'd like to integrate it into an Outlook add-in with a Microsoft Graph API webhook to automatically detect new emails rather than using a script to continuously check.
Some challenges we faced:
- Figuring out how to train the model so it consistently replies professionally but not too bot like
- Separating each question to generate a response to, and recombining the responses into one email.
- Working with unfamiliar tools
- Setting up the environment
- Navigating Microsoft tools / documentation