NYC EdTech Hackathon

NYC EdTech Hackathon

2025

I contributed to end-to-end UX design for GradeSnap, an AI-powered grading tool built to integrate with Google Classroom, within a five-week hackathon sprint. As a former K-12 educator, my firsthand classroom experience directly shaped the team's research approach and key design decisions throughout the project.

Our solution reduced teacher grading time from 20 minutes to 4 minutes per assignment, a 5x efficiency improvement validated through usability testing. We placed 5th overall and were featured on the NYC EdTech Hackathon website.

Team: Cross-functional (engineering, UX, PM)
Role: Design Research & Interaction Design
Timeline: February–March 2025

Project image

Framing the Opportunity

Grading Is a Systemic Strain

Teachers report working ~54 hours per week.
Writing assignments, while critical, are among the most time-intensive tasks.

In interviews, one teacher shared:

“Sometimes I spend four hours grading just one class. I love teaching, but I’m exhausted.”

Existing tools organize assignments, but don’t meaningfully support feedback workflows.

We asked:

How might AI assist teachers without replacing their judgment?

Project image

Research Sprint

Understanding Workflow & Friction

In Week 1, we conducted:

  • 5 semi-structured interviews with K–12 teachers

  • Competitive analysis of Gradescope, Crowdmark, and GradeMate

We mapped grading workflows and tensions:

  • Feedback fatigue

  • Cognitive load during rubric scoring

  • Weak Google Classroom integration across existing tools

Project image

Translating Insight Into Concept

GradeSnap

We prototyped an AI-assisted grading flow:

  1. Sync assignment from Google Classroom

  2. AI generates rubric-aligned draft feedback

  3. Teacher reviews, edits, approves

  4. Scores + feedback return in one click

The teacher remains the decision-maker.
AI functions as cognitive support.

Project image
Project image
Project image
Project image

Final Product

What This Reinforced

  • Speed doesn’t eliminate rigor.

  • AI tools must respect professional expertise

  • Systems thinking applies even in compressed timelines

This project reinforced my interest in designing tools that reduce cognitive load without removing human agency.