Skip to content
TOLGA İLERİ

Home/Work/ZihinLAB

Project / 022026AI-Powered Education Platform

ZihinLAB

A multi-role AI education platform combining personal AI tutoring, visual question understanding, digital publishing, learning analytics, study planning and gamified learning.

Role
Full-Stack Development / Product Engineering / Applied AI
Type
AI
Stack
Next.js / TypeScript / PostgreSQL / Supabase / DeepSeek / Gemini 2.5 Flash
Status
Live Product

Overview

ZihinLAB is a full-scale AI education platform designed for students, teachers, parents and schools. It combines AI-assisted tutoring, image and PDF question understanding, study tools, digital question banks, practice exams, planning, institutional dashboards and gamified learning within a unified product ecosystem.

Role
Full-Stack Development / Product Engineering / Applied AI
Year
2026
Type
AI
Stack
Next.js / TypeScript / PostgreSQL / Supabase / DeepSeek / Gemini 2.5 Flash

01 / Context

K–12 and YKS-oriented learning needs more than isolated homework tools. Students, teachers, parents and schools need role-aware workflows that can handle tutoring, content, assessment, planning and progress together.

02 / Challenge

The platform brings together very different educational workflows — AI tutoring, content publishing, assessments, planning, school administration and gamification — while maintaining role-specific experiences for students, teachers, parents and administrators.

03 / Solution

ZihinLAB uses a modular product architecture with role-aware workflows, PostgreSQL-backed educational data and routed AI services for text reasoning and vision/OCR tasks.

Responsibilities

  1. 01Designed and developed the multi-role education platform.
  2. 02Built AI tutoring and visual question-processing workflows.
  3. 03Developed digital publishing, question-bank and exam systems.
  4. 04Implemented student planning, progress and gamification systems.
  5. 05Built school, teacher, parent and administrative workflows.
  6. 06Integrated AI model workflows for text reasoning, OCR and structured educational outputs.

Key areas

  • Personal AI Teacher
  • Image and PDF question understanding
  • DeepSeek-powered AI workflows
  • Gemini vision/OCR integration
  • Digital publishing system
  • Question banks and practice exams
  • Study planning
  • Multi-role school platform
  • Gamification and AI Adventure Academy
  • Student progress and reporting

Technology

Architecture

  • Multi-role education platform for students, teachers, parents and schools
  • Routed AI workflows for text reasoning and vision/OCR

Frontend

  • Next.js
  • TypeScript

Backend

  • TypeScript

Database

  • PostgreSQL
  • Prisma
  • Supabase

AI

  • DeepSeek v4 Flash for text reasoning
  • Gemini 2.5 Flash for OCR and visual extraction
  • Structured educational outputs

Infra

  • Docker

Integrations

  • Supabase
  • iyzico