ApplyGrad
Helping experts make
better decisions—not faster ones.
Research Lead & Client Communications Lead
8-month client engagement with Carnegie Mellon School of Computer Science
The Carnegie Mellon School of Computer Science partnered with our team to redesign Apply Grad, a graduate admissions platform supporting faculty as they evaluate thousands of applications each admissions cycle. As Research Lead and Client Communications Lead, I defined the research strategy, led discovery and synthesis, and partnered closely with stakeholders to understand how admissions decisions were actually made.
What began as a usability challenge ultimately became a lesson in designing for human judgment—revealing that fairness, confidence, and collaboration mattered far more than efficiency alone.
Graduate admissions isn't a checklist.
It's a conversation.
Faculty members evaluate applicants across academic achievement, research experience, portfolios, recommendation letters, and personal statements while also balancing the needs of an incoming cohort.
Every reviewer brings different expertise.
Every decision is subjective.
Every decision matters.
The platform supporting that work often got in the way of it.
One admissions officer summarized the experience better than we ever could:
"I’m looking for reasons to admit a person, not reject them. It’s really exciting to see and learn about these applicants, but it’s just such a long process.
It's really draining."
Before proposing solutions, we needed to understand how admissions decisions were actually made—not just how the software supported them. We spent time with admissions officers observing their workflows, understanding how they evaluated applicants, and identifying where the current experience created unnecessary friction.
Research Activities
Contextual Inquiry • Stakeholder Interviews • Journey Mapping • Collaborative Synthesis • Usability Testing • Concept Evaluation
To understand where friction existed, we mapped the current admissions workflow from initial application review through collaborative decision-making. Rather than beginning with interface screens, we focused on understanding the work itself.
Applications weren't reviewed in isolation. Faculty routinely discussed candidates, advocated for applicants, and worked together to shape an incoming cohort.
Each admissions officer brought unique expertise and evaluation criteria, making flexibility just as important as consistency.
Reviewers constantly switched between systems, external notes, spreadsheets, and discussion, creating unnecessary overhead during an already demanding process.
"Each reviewer is probably looking at like, 60 to 80 [applications] and we can spread it out over several weeks. So, you know, for me to remember somebody I reviewed in week one and I'm now in week eight... you know, it can start to get blurred a little bit."
We mapped the admissions journey from application review through final decision to understand where the experience was creating unnecessary cognitive load.
Three recurring friction points emerged.
Information lived everywhere.
Reviewers relied on multiple windows, spreadsheets, and external notes to complete a single evaluation.
Collaboration happened outside the platform.
Many conversations and annotations occurred through email or personal documents rather than inside ApplyGrad.
Administrative work distracted from meaningful evaluation.
Admissions officers wanted to spend their time understanding applicants—not managing the mechanics of the process.
At first, these findings seemed to point toward a familiar conclusion.
Reduce friction.
Increase efficiency.
Simplify the workflow.
But something didn't add up.
The people using the system weren't asking to work faster.
They were asking to make better decisions.
Halfway through the project, we presented our research and early concepts to stakeholders. The response was mixed.
While our work reflected many of the workflow challenges we'd observed, it didn't resonate with the people we were designing for. We had become focused on improving the system.
We hadn't fully understood what success meant. That realization forced us to step back and ask a different question.
What does success actually look like for an admissions officer?
The answer wasn't efficiency.
Through continued interviews and observation, we realized admissions officers weren't trying to review more applications. They were trying to make fairer decisions. The administrative work surrounding admissions was exhausting—but evaluating applicants wasn't.
What frustrated them wasn't the decision-making. It was everything that distracted from it.
"I don't think efficiency is necessarily the goal. I think the discussion was more in-depth, maybe slightly more in depth than it might be typically. I think that's a good thing. So even if the conversations are a bit more extended than it might be... just slightly... I think that's all for the better."
Instead of asking "How can we help admissions officers work faster?" we started asking
"How can we help them make thoughtful decisions with greater confidence?"
That single shift changed every design decision that followed.
We stopped designing for...
We started designing for...
Speed
Confidence
Automation
Human judgment
Task completion
Thoughtful evaluation
Individual evaluations
Collaborative decision-making
These principles became the lens through which every design decision was evaluated.
Once we understood that admissions officers weren't optimizing for speed, our design priorities changed. Rather than redesigning the interface around efficiency alone, we focused on creating tools that supported thoughtful evaluation, reduced unnecessary cognitive burden, and strengthened collaboration throughout the admissions process.
Research showed that reviewers frequently relied on external notes throughout the admissions process.
Designing persistent annotations allowed them to capture context directly within the platform, reducing administrative overhead without disrupting individual review styles.
Research showed that admissions officers spent a significant amount of mental energy managing the review process itself. Navigating dozens of applicants, tracking review progress, and comparing candidates across multiple criteria often required unnecessary effort before meaningful evaluation could even begin.
Rather than simplifying the admissions process, we focused on making complex information easier to understand and navigate. The interface reduced unnecessary mental overhead while preserving the depth and flexibility required for expert evaluation.
Admissions decisions extended beyond individual reviews. Faculty members brought different expertise, priorities, and evaluation styles to each applicant, making discussion and consensus-building a critical part of the process. Rather than averaging opinions into a single score, we designed the experience to preserve each reviewer's perspective while making it easier to compare feedback, identify patterns, and understand the reasoning behind every evaluation.
Our goal was to make individual perspectives visible so committee discussions could begin with shared context instead of scattered notes.
The most valuable outcome of this project wasn't the interface—it was learning how quickly an understandable assumption can shape an entire design direction. We started by optimizing for efficiency, then discovered that the people using the system valued something very different: the time and space to make thoughtful, fair decisions.
That experience reinforced something I carry into my work today: research isn't just about validating a solution. Sometimes its most important job is showing us that we're solving the wrong problem.
Technology should support human judgment—not replace it.
