Back to Projects
2025 – Present·PRISM
Full-Stack Developer, Beta at Scholars Education

PRISM.

The operating system for tutoring centers.

Interactive Preview

See It in Action

By the Numbers

7

App Modules

Sessions, check-in, messaging, homework, reports, admin, logbook

4

User Roles

Super Admin, Center Admin, Tutor, Student (no-login QR)

End-to-End

AI Pipeline

Voice → AI → Manager Review → Parent Email with homework

Overview

I identified an operational gap at Scholars Education and built PRISM from scratch. It's currently in beta testing with the team providing feedback as we refine the platform. PRISM manages session workflows, parent communication, AI homework generation, and team messaging. The entire pipeline from tutor voice notes to polished parent emails is automated.

Architecture

Pattern

Next.js App Router + Server Components

DI Strategy

Server-side role middleware with React cache()

AI Integrations

Google Gemini 1.5 Pro

Backend

Supabase (PostgreSQL) with RLS + Resend email

Testing

Playwright E2E

Key Features

What I Built

AI Session Notes Pipeline

Tutors speak or type raw notes → Gemini generates professional parent-facing summary → Manager reviews/rewrites → Approved note triggers email with AI-generated homework.

QR Check-In System

Students scan unique QR codes (no login required). Multi-step animated form captures grades, upcoming tests, homework status, and session rating.

Real-Time Team Messaging

Slack-like messaging with channels, DMs, announcement-only modes, and Supabase Realtime subscriptions.

AI Homework Generation

Grade-appropriate practice problems generated by Gemini. Embedded directly in parent emails with hints and difficulty tags.

Multi-Tenant Architecture

Centers as first-class entities. All queries scoped by center_id. Per-center settings, user management, and data isolation.

Performance

Optimizations

Server-side rendering with React cache() for auth

Supabase Realtime for instant messaging

Optimistic UI updates with Sonner toasts

Voice input via Web Speech API for frictionless note-taking

Tech Stack

Next.jsTypeScriptGoogle GeminiSupabaseResendReact EmailThree.jsFramer Motion

Interested in learning more?

Check out the live project or get in touch to discuss the technical details.