Shipped product Live demo available

Exam Preparation

Adaptive assessments and AI-powered insight for K-12 schools

Request Demo View Live Site Explore features ↓

Multi-tenant ready
Production deployed
Web + mobile

Built by Toshi

Case study inside

Overview

Exam Preparation (internally ExamPrep / the Adaptive Assessment Analyzer) is a multi-tenant, B2B2C EdTech platform for K-12 schools, deployed live at examprep.sentinal-ai.in. It serves four roles from one backend:

Students take tests and review results
Teachers author and assign custom tests
School admins manage their school (teachers, classes, students, billing)
Super admin owns the central question bank, schools, and platform-wide analytics

The product runs two assessment modes. Adaptive tests draw from a platform-owned question bank where difficulty auto-adjusts per answer. Custom tests are teacher-authored, fixed-question sets assigned to specific classes or sections. After every test, students receive rule-based analytics plus a Groq-LLM narrative report covering strengths, weaknesses, and recommendations.

Multi-tenancy is enforced server-side: a school admin or teacher can never read or mutate another school's data.

Tech stack: FastAPI + SQLAlchemy on Postgres 16 (backend), Next.js 16 + React 19 App Router (web), and Expo SDK 56 + Expo Router + NativeWind v5 (mobile). AI reports run on Groq; email via Resend; payments via Razorpay; self-hosted on a Contabo VPS behind Caddy via Docker Compose.

Problem Statement

Indian K-12 schools largely assess students with one-size-fits-all paper tests that reveal a final score but little about why a student struggles or which topics to revisit. Static tests fail to calibrate to ability — strong students coast through easy items while weaker students stall on questions far above their level.

Exam Preparation closes that gap on two fronts:

Adaptive testing that tunes difficulty to each student in real time
Plain-English AI reports that translate accuracy-by-topic numbers into guidance a student or parent can act on

For schools, the goal is a managed, multi-tenant assessment system with per-class analytics — not a generic quiz tool.

Case Study

Architecture

A single FastAPI backend exposes routers per concern — auth, questions, adaptive, custom, reports, analytics, schools, billing — over a ~30-table Postgres model spanning identity, the question bank, the assessment engine, results/AI, and billing domains.

Adaptive engine

The adaptive engine drives a difficulty ladder with start, next, current-question, and submit endpoints. An anti-cheat auto-flag watches for tab-switches and implausibly fast answer times — speed alone never flags a session unless paired with tab-switching.

Role dashboards & analytics

Scoped analytics for students, teachers, school admins, and platform super admins. Teachers are limited to their own sections via TeacherSection assignments.

Billing & trial

Students get 3 free adaptive attempts/month via a scheduled credit job plus paid packs through Razorpay. Schools subscribe to Starter / Pro / Enterprise plans with server-side caps and GST invoice PDFs.

Notable engineering decisions

Web auth uses httpOnly cookies with proxy middleware — no localStorage tokens
Mobile uses Bearer tokens in expo-secure-store with refresh-token rotation and reuse detection
Idempotent SQL migrations on startup (no Alembic) with demo dataset seeding
Strict web + mobile parity for student/teacher features; super-admin tooling web-only

Screenshots

Marketing landing page — Marketing home — cyan-brand landing with persona-switching dashboard mockups.

Login with demo accounts — Login screen with one-click demo accounts for each role.

School admin dashboard — manage teachers, classes, students, and billing within one tenant.

Super admin dashboard — central question bank, schools, users, and platform-wide analytics.

Interested in Exam Preparation?

Talk to our team about deployment, customization, or a demo for your organization.