HoyaMind
An adaptive burnout check-in tool for Georgetown students
AI
adaptive testing (CAT)
learning analytics
what if wellness check-ins were adaptive, fast, and actually led students to the help they need?
Graduate students experience high levels of stress, burnout, and anxiety. At the same time, free wellness resources on campus are underutilized.
HoyaMind is a speculative product concept that uses a computerized adaptive test (CAT) model to assess a student’s level of burnout and recommend support services at Georgetown University. It explores how learning analytics, AI, and design can support student mental health in scalable, human-centered ways.
seeing the gap
While I was a graduate student at Georgetown, I started noticing just how many wellness resources the university offers — yoga, swimming, tennis, meditation rooms, counseling services, group therapy, individual emergency sessions, academic coaching, and more. It’s a wide and generous ecosystem of support.
But what stood out to me was how many of my peers weren’t aware these resources existed. I first heard about them from second-year students in my program, and realized that this kind of knowledge often spread informally — through conversation, chance, and shared experience.
In fast-paced, one- or two-year graduate programs, students are often focused on classes, internships, and deadlines. It’s easy to miss what’s available until someone happens to mention it. And that stayed with me.
One day during finals, a classmate said, “I don’t even know how I’m doing, I just know I’m tired all the time.” That quiet moment sparked something.
What if there was a way to help students check in with themselves and gently connect them to the resources already around them?
That’s when the idea for HoyaMind began to take shape.
The Problem
Graduate students report high rates of burnout and mental health strain
Many students don't know about campus resources or how to access them
Generic surveys are often long, not tailored, and have low engagement
Students need a quick, sensitive, personalized way to check in with themselves
The Idea
We imagined a mobile app that asked students a few simple questions and gave them personalized, friendly guidance. Not a diagnosis — just a check-in.
We decided to build it using computerized adaptive testing (CAT) — the same logic behind adaptive learning — but in service of mental health.
And we called it HoyaMind.
How We Built It
Adaptive Testing with AI. We used ChatGPT to generate a bank of True/False questions at three difficulty levels. These questions were refined to assess burnout symptoms while staying emotionally neutral and appropriate.
I engineered the prompts and then evaluated item performance using:
Item Difficulty (b)
Item Discrimination (a)
Reliability (Cronbach’s α = 0.95)
Burnout Score Interpretation
We used a theta score to place students in one of three burnout zones:
Low Burnout: Encouraging message + maintenance tips
Moderate Burnout: Gentle support + free resources
High Burnout: Supportive tone + direct links to mental health services
What We Created
HoyaMind became a proof-of-concept for how adaptive testing, learning analytics, and empathetic design can come together to support student well-being.
Students answer 5–10 questions. The system updates their burnout score after each one and stops when enough precision is reached. At the end, students see:
A burnout level
A message written with care
Campus resources tailored to their needs
What I Learned
This wasn’t just a data project or a wellness app. It was a systems design challenge.
I learned how to:
Use psychometric tools to evaluate question quality and test reliability
Design with AI ethically and critically - not just for efficiency, but for humanity
Create a product experience that balances analytics, usability, and emotional tone
Mostly, I learned that assessment is never just technical - it’s also cultural, emotional, and relational. And if we treat it that way, it can become a powerful design tool for care.
