Skip to main content
Live Demo
SAFE-Bridge — case study by Faisal Abbas

SAFE-Bridge

Next.jsTypeScriptExpressPostgreSQLPrismaOpenAITailwind CSSZustand

About This Project

SAFE-Bridge is an AI-powered Academic Engagement Intelligence Platform built entirely solo as a student project. It transforms 100% anonymous student feedback into actionable insights for educators using GPT-4o-mini for real-time text classification, a custom Psychological Safety Index (PSI), and predictive analytics with linear regression models.

The Problem

Students rarely give honest feedback for fear of exposure, and educators are left without the actionable, data-backed insight they need to improve engagement.

The Solution

An AI Academic Engagement platform — built entirely solo — that collects 100% anonymous feedback, classifies it in real time with GPT-4o-mini, scores it with a custom Psychological Safety Index, and forecasts trends with predictive analytics.

Key Features

GPT-4o-mini powered real-time feedback classification
Custom Psychological Safety Index (PSI) algorithm
Predictive analytics with linear regression models
14 database models with Prisma ORM
12 backend services with Express.js
30+ React components with Zustand state management
3-language support (English, Urdu, Arabic RTL)
100% anonymous feedback collection
Statistical outlier detection using IQR method
Department-level analytics and reporting

Impact & Results

Solo-built: 14 database models, 12 services, 30+ components
Real-time GPT-4o-mini feedback classification
Custom Psychological Safety Index + predictive analytics
3-language support (English, Urdu, Arabic RTL)
Business Value

Turns anonymous sentiment into a measurable, trackable signal institutions can act on — and proves the ability to ship an enterprise-grade platform independently.

Technical Challenge

Building an enterprise-grade platform solo — from 14 database models to 12 backend services to 30+ frontend components — required careful architecture planning. The custom PSI algorithm needed extensive research into psychological safety measurement frameworks.

Tech Stack

Next.js
TypeScript
Express
PostgreSQL
Prisma
OpenAI
Tailwind CSS
Zustand
Visit Live Demo