UniWizard.

The starting point.

The goal was a guided way for a student to say what matters, and to get personalized school and degree matches back.

What we built.

A guided questionnaire, not a search box

Students move through a step-by-step questionnaire about their priorities. The answers feed a results page and a personal dashboard, with a page for each university, school, and degree.

An AI engine that profiles every university

• A Python engine uses Google Gemini 2.5 Flash to score about 100 Canadian universities on 23 attributes: academic reputation, admission difficulty, cost, co-op and internship opportunities, social life, campus, city size, weather, alumni network, teaching quality, and others. Each score sits on a scale from −1 to 1.
• Language models vary from run to run, so the engine samples each score in batches until the result stabilizes, at a standard error of 0.02 or below.

A production-minded foundation

• Sign-up and login with Auth0, with users synced to a Spring Boot API backed by MySQL.
• A typed React frontend with Zod validation, TanStack Query for data caching, and a component system on shadcn/ui.
• Unit and end-to-end tests with Vitest, Playwright, and JUnit, and continuous integration on GitHub Actions.

How the matching works.

A student’s answers become a weighted preference profile. Universities are ranked by how closely their attribute scores fit that profile, using a weighted distance. Two students can receive very different lists, and both lists can be right.