Inspiration
DTU handles thousands of student queries but most get delayed or misplaced due to manual routing. We wanted to build an AI system that ensures every student’s concern reaches the right department instantly, reducing frustration and improving transparency.
What it does
UniHelp AI acts as an AI-powered helpdesk for universities.
Students submit issues in plain language.
An NLP model analyzes the message and auto-sorts it to the correct department (IT, Hostel, Academics, Admin, etc.).
The system generates a tracking ID and shows the complaint’s real-time status.
Admins can view analytics on issue volume, response time, and sentiment trends.
How we built it
Frontend: HTML, CSS, JavaScript (clean student dashboard + complaint tracker).
Backend: Node.js and Express.js with MongoDB for issue storage and tracking.
AI Layer: Google Gemini API linked which act as a llm
Integration: REST API connects frontend with the AI classifier for real-time auto-routing.
Challenges we ran into
Creating accurate training data for varied student language styles.
Handling multi-intent complaints (e.g., hostel + Wi-Fi in one query).
Integrating Python AI model with Node.js backend efficiently.
Ensuring security and privacy for complaint data.
Accomplishments that we're proud of
Built a functional AI pipeline that routes queries automatically with over 90% accuracy.
Designed a modern, responsive dashboard for both students and admins.
Internally tested with simulated DTU use cases — successfully handled 200+ mock queries.
Scalable design ready for real university deployment.
What we learned
Importance of structured backend design when integrating AI models.
Building an AI product that balances usability, accuracy, and trust.
What's next for 'hackdays_unihelp
Add Generative AI chatbot to answer common university FAQs.
Implement sentiment-based prioritization for urgent complaints.
Multi-language query support for inclusivity.
Deploy as a web + mobile app for campus-wide adoption.
Log in or sign up for Devpost to join the conversation.