VidyaSetu
Full-stack AI assessment & analytics
Role
Full-Stack Developer
Timeline
2026
Stack
Next.js
Supabase
FastAPI
Langchain
Groq
Overview
An adaptive study platform generating quizzes from NCERT content, evaluating subjective answers via AI, and providing a robust student dashboard with analytics.
The Challenge & Architecture
VidyaSetu was built to democratize high-quality education by transforming standard NCERT textbooks into highly interactive, AI-driven study sessions.
The platform relies on a Next.js (React Server Components) frontend powered by Supabase (PostgreSQL + pgvector) for database and authentication.
The heavy lifting for subjective evaluation and document querying is offloaded to a standalone Python FastAPI AI backend that orchestrates Sentence Transformers and Langchain (with Groq models).
System Architecture
Next.js UIReact Server Components
Python BackendFastAPI & Langchain
AI ModelsGroq & Transformers
SupabasePostgreSQL + pgvector