← ALL WORK
MISSION INTEL // April 2026

A diagnostic that tells a language barrier apart from a real knowledge gap for English Language Learners.

3rdVibra ATL 2026 (SHPE, GSU)Frontend Lead
  • Next.js 16
  • React 19
  • Web Speech API
  • JWT
  • SQLite
  • OpenAI

May take ~30s to wake

Saber interface, view 1

THE PROBLEM

When an English Language Learner misses a question, a teacher cannot tell whether the student did not understand the concept or just did not understand the English. Those two problems need opposite fixes, and lumping them together holds kids back.

THE BUILD

I led the frontend on a diagnostic that separates the language barrier from the knowledge gap. A teacher uploads scores, and the student retakes missed questions in Spanish. Get it right in their first language and it was a language issue; still miss it and it is a real knowledge gap. I built it in Next.js 16 and React 19 with Web Speech API voice input and JWT auth, over an OpenAI-backed classify and translate layer.

THE RESULT

Saber placed 3rd at Vibra ATL 2026, the SHPE hackathon at Georgia State. It is deployed on Render's free tier, so the live link takes about 30 seconds to wake from a cold start.