Pariksham – AI Powered Smart Examination System

• 

• 

Pariksham – AI-Powered Online Examination SaaS Platform Inspiration

The idea for Pariksham (परीक्षा – "Examination" in Hindi) came from witnessing the challenges educational institutions face in conducting secure, scalable online exams. During the pandemic, many schools and coaching institutes struggled with cheating, proctoring, and exam management. I wanted to create a multi-tenant, AI-enabled SaaS platform that allows institutes to manage exams efficiently, provides real-time analytics, and empowers students with personalized learning insights.

What it does

Pariksham is a full-featured online examination platform designed for students, institute admins, and master admins:

For Students: Secure login, advanced exam console, real-time proctoring, instant results, AI-generated practice questions, and personalized study recommendations.

For Institute Admins: Manage students, exams, and questions; AI question generation; performance analytics; branding; and subscription management.

For Master Admins: Multi-tenant oversight, payment verification, platform configuration, and analytics dashboard.

AI Capabilities: Weak topic analysis, study recommendations, practice question generation, and difficulty analysis using Google Gemini AI.

How we built it

Frontend: Next.js 16 (App Router) + React 19 + TypeScript

Styling: Tailwind CSS v4 + Shadcn/UI components

Backend: Firebase (Auth, Firestore, Cloud Storage)

AI: Google Gemini 1.5 Flash for question generation, practice mode, and recommendations

Payments: UPI manual submission

Email & Notifications: Nodemailer (SMTP)

Data Handling: Papa Parse for bulk CSV upload

Architecture Highlights:

Firestore collections for users, institutes, students, exams, questions, responses, results, warnings, payments, subjects, and topics.

API routes for AI tasks: /generate-questions, /generate-practice, /analyze-performance.

Role-based access control with Master Admin, Institute Admin, and Student.

Challenges we ran into

Proctoring & Anti-Cheating: Implementing real-time tab-switch detection, keyboard shortcuts, fullscreen enforcement, and auto-submission was complex.

Multi-Tenant Data Isolation: Ensuring institute-specific data separation and security in Firestore required precise rules.

AI Integration: Generating high-quality questions and recommendations while maintaining performance and schema compatibility.

Scalability & UX: Designing a responsive UI for thousands of students while keeping the platform fast and intuitive.

Accomplishments that we're proud of

Successfully integrated AI-powered question generation and personalized learning recommendations.

Built a production-ready, multi-tenant SaaS platform serving schools and coaching institutes.

Achieved advanced proctoring features that prevent cheating without intrusive monitoring.

Developed a practice mode with unlimited AI-generated questions per student.

Implemented performance analytics for institutes to identify weak topics and difficult questions.

What we learned

Deep understanding of Next.js 16 App Router, React 19, and Tailwind CSS for scalable frontend development.

Practical experience with Firebase Auth, Firestore, and Cloud Storage, including security rules for multi-tenant apps.

How to integrate AI APIs (Google Gemini) for educational content generation.

Best practices for role-based access control, responsive design, and real-time features.

Handling user experience challenges while maintaining exam security.

Built With

• ai
• cloud
• csv
• email:
• firestore
• flash
• gemini
• google
• languages-&-frameworks:-javascript
• manual
• next.js-16
• nodemailer
• papa
• payments:
• react-19-styling-&-ui:-tailwind-css-v4
• shadcn/ui-database-&-backend:-firebase-auth
• storage
• submission
• typescript
• upi