PulseLab

Inspiration

College campuses are full of opportunities — scholarships, jobs, events, research — but most students miss them.

Not because they don’t care. Because the system is broken.

Information is scattered across portals, emails, PDFs, and word-of-mouth. There’s no layer that connects:

who you are
what’s happening
what you should do right now

We kept asking: What if your campus worked like a real-time intelligence system instead of a collection of static dashboards?

That’s how PULSE was designed, not as another app, but as a decision layer for student life.

What it does

PULSE is a real-time, agent-driven campus intelligence system.

It continuously monitors campus signals and transforms them into personalized, actionable opportunities.

For students:

Matches scholarships and pre-fills applications
Surfaces on-campus jobs aligned with skills
Recommends relevant events and auto-schedules them
Tracks deadlines before they become problems
Calculates a Campus IQ Score (engagement, momentum, opportunity usage)

For faculty:

Flags at-risk students early
Drafts context-aware intervention emails
Surfaces research + funding opportunities
Measures student engagement impact

Instead of reacting to missed opportunities, PULSE acts before they’re lost.

How we built it

We designed PULSE as a modular agent orchestration system layered on top of a campus AI platform.

Core Architecture:

Agent Orchestrator – coordinates multiple task-specific agents
Opportunity Agents – scholarships, jobs, events, deadlines
Matching Engine – ranks opportunities based on relevance + urgency
Action Layer – executes tasks (apply, RSVP, schedule, notify)
Scoring Engine – computes Campus IQ + engagement metrics

Backend:

Node.js / TypeScript
Modular API layer (/pulse system)
Agent routing + orchestration logic

Frontend:

Next.js (App Router)
Dynamic dashboard UI
Real-time actionable cards

Intelligence Layer:

LLM-powered:
- Opportunity extraction
- Application drafting
- Email generation
- Structured + heuristic matching for deterministic reliability

We intentionally built it as a non-invasive extension — plug-and-play into existing campus ecosystems.

Challenges we ran into

One of our biggest challenges was that TerpAI is designed as a standalone interface for UMD students, not as a developer-ready platform with public APIs that could be directly plugged into another application. Since PULSE depended on intelligence and workflows that needed to connect seamlessly into a larger system, this created a major integration barrier.

To solve this, we had to think beyond simple frontend prototyping. Instead of relying on nonexistent native integrations, we designed and developed an API layer from scratch that could interface with our system architecture and make PULSE operational. That meant building the connection logic ourselves, structuring request/response flows, and creating a clean way for PULSE to use TerpAI-driven capabilities inside a broader campus intelligence experience.

Accomplishments that we're proud of

Building an API Architure to Export TerpAI API from Stratch, Agents Directly from your TerpAI dashboard to Anywhere you want
Built a multi-agent campus intelligence system in hackathon time
Designed a new paradigm: campus as an operating system
Created actionable flows, not just insights (apply, schedule, respond)
Introduced Campus IQ Score to quantify student momentum
Delivered dual impact:
- Student success
- Faculty intervention

Most importantly:

We didn’t build another dashboard.
We built a system that changes behavior in real time.

What we learned

Insight without action is useless:-- Systems must do, not just inform
Personalization is everything:-- Relevance drives engagement more than features
Agent systems > single AI models:-- Coordination creates real-world utility
Timing matters more than availability:-- The right opportunity at the wrong time is still a miss