RU on Campus

Agentic AI platform for Rutgers Student helping with various tasks

Comment

🎓 RU On campus — Your Personalized Campus Companion

“What if your university life had a voice — one that understands you, plans for you, and takes care of the little things so you can focus on the big ones?”

💡 Inspiration

College life is an incredible mix of opportunities and chaos. Between finding study rooms, grabbing coffee before lectures, rushing to hackathon workshops, and figuring out which bus takes you downtown — we realized how much mental load a student carries every single day.

Our team kept joking during all-nighters:

“If only there was an assistant that could book our study room, order a pizza, and remind us about the robotics event — all by voice command!”

That became our eureka moment. Instead of just another chatbot, we envisioned something smarter, proactive, and agentic — a personal campus concierge that listens, reasons, plans, and acts autonomously.

That’s how RU On campus was born — an intelligent, voice-driven assistant designed to make campus life simpler, smarter, and more connected.

🧠 What We Built

Our project simulates a multi-agent AI system that handles all major aspects of student life. Even without real university APIs, we built mock datasets and APIs to bring the experience to life — showcasing reasoning, planning, and real-time voice interaction.

🎯 Core Modules

1️⃣ Study Buddy Concierge

A mock booking system lets students reserve study rooms using natural voice.

“Reserve Room 204 for me from 3 to 5 PM.” The AI checks room availability from a JSON file, confirms the slot, and replies with a lifelike ElevenLabs voice: “I’ve reserved Room 204 for you from 3–5 PM. Happy studying!”

2️⃣ Campus Event Recommender

We created a mock event list — hackathons, movie nights, club meets, sports.

“What can I do tonight for 2 hours?” The AI filters upcoming events, recommends the best match, and guides the user using Google Maps. “The Robotics Club is hosting a demo at 6 PM in Hall A. I can guide you there!”

3️⃣ Transit & Navigation Concierge

This module helps students travel across campus or to downtown using public transit. It integrates with Google Maps or a mocked /routes API.

“How do I get from Livingston Dorms to Hill Center?” The AI returns the route, departure times, and even narrates directions using ElevenLabs voice — or a fun “pirate” accent for humor.

4️⃣Daily Life Automator

This one acts like a life organizer — connecting with Google Calendar (or mock schedules).

“Block 2 hours to study, then remind me to call mom at 7 PM.” It sets reminders, schedules study sessions, and confirms with a natural voice: “All set! Your study time and gym slot are booked, and I’ll remind you to call mom at 7.”

🛠️ Tech Stack

Category Tools & Frameworks
Frontend React(for interactive UI & voice input)
Backend FastAPI(Mock APIs for rooms, routes, food, and events)
AI Core Gemini (Agent reasoning & function-calling)
Voice ElevenLabs (Text-to-speech & personality voices)
Orchestration LangChain-style agent routing
Data JSON datasets (for mock API simulation)

🧩 Implementation Strategy

We knew we couldn’t depend on real campus APIs, so we designed a hybrid approach:

  • Mock APIs: Built with Flask returning JSON responses for /rooms, /events, /menu, and /routes.
  • AI Reasoning Layer: A Gemini-based logic layer determines which agent to trigger depending on intent.
  • Voice System: ElevenLabs voice synthesis adds realism and engagement.

Each module can work independently or collaboratively

🧠 What We Learned

Building the Campus Concierge taught us much more than API calls or TTS integration. We learned how to design for autonomy — how agents can reason, prioritize, and act, rather than just respond.

Key learnings:

  • How to combine LLMs with structured logic for hybrid AI.
  • Handling intent detection, task routing, and multi-agent coordination.
  • Working with mock APIs to simulate a real ecosystem.
  • Using voice synthesis to humanize the AI experience.
  • Designing modular architecture for plug-and-play scalability.

⚡ Challenges We Faced

  • 🚫 Lack of campus APIs: We had to simulate entire datasets — rooms, routes, and menus — but it made us creative.

* 🤖 Complex agent orchestration: Ensuring the right agent was called at the right time required prompt tuning and reasoning constraints.

📈 Impact and Future Vision

We see RU On campus as more than a hackathon project — it’s a glimpse into the future of smart campuses. With real API integration, it could:

  • Manage student schedules dynamically.
  • Recommend personalized events based on past preferences.
  • Handle transit updates in real-time.
  • Even connect to campus facilities for real bookings.

Long-term, we envision it evolving into an AI-powered campus ecosystem that understands every student’s context — learning patterns, habits, and routines — to make university life efficient and delightful.

🏁 Conclusion

RU On campus isn’t just a voice assistant — it’s an agentic AI ecosystem that learns, acts, and assists like a real companion. It bridges the gap between convenience and intelligence, showing how AI can transform university life from scattered tasks to seamless experiences.

As we move forward, we plan to integrate real Rutgers APIs, add RAG-based personalization, and expand the assistant beyond campuses — creating a universal, context-aware lifestyle companion for students everywhere.

Built With

Share this project:

Updates