Skip to content
 
 

Latest commit

 

History

79 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hyper Personalized Tutor - The Expert Network for Learning

Mustaqbil Hackathon 2026 | Track 1: Generative AI & Agents | Challenge B

Team Agentex

  • Huzair Ahmed Khan
  • Muhammad Fasih
  • Anas Ahmed Shaikh
  • Hashir Hussain

The Problem

Traditional educational tools offer a "one-size-fits-all" approach that fails to adapt when a student is confused. This leads to "confusion loops" where learners disengage because the material doesn't pivot to their specific needs or learning style.

Our Solution: The 10,000+ Expert Network

Hyper Personalized Tutor isn't just another AI tutor - it's an intelligent network of 10,000+ specialized AI experts, each designed to eliminate confusion through precision teaching tailored to your exact level of understanding.

How It Works

Intelligent Expert Matching When you ask a question, our system doesn't give you a generic response. Instead, it:

  1. Analyzes your confusion level and learning context
  2. Matches you with the perfect specialized expert from our network
  3. Deploys that expert to explain the concept in the way YOU need

Adaptive Confusion Detection Each expert in our network is equipped with advanced confusion detection that:

  • Identifies exactly where your understanding breaks down
  • Adjusts explanation depth in real-time based on your responses
  • Pivots teaching strategies when confusion persists
  • Ensures complete conceptual clarity before moving forward

Multi-Modal Explanations Our experts don't just use text - they leverage every medium necessary to ensure understanding:

  • Text Explanations: Clear, concise answers at your comprehension level
  • Mermaid Diagrams: Visual flowcharts, graphs, and concept maps
  • Image Generation: Custom visuals and illustrations for complex concepts
  • Interactive Quizzes: Test understanding and reinforce learning
  • Step-by-Step Breakdowns: Progressive complexity building
  • Code Examples: Live, executable demonstrations (when relevant)
  • Analogies & Metaphors: Relatable comparisons tailored to your background

100% Accurate, Verified Information Every expert in our network is backed by:

  • Advanced RAG (Retrieval-Augmented Generation): Domain-specific knowledge bases ensuring factual accuracy
  • Real-Time Web Search: Up-to-date information when needed
  • Cross-Verification: Multiple knowledge sources validate each explanation
  • Source Attribution: Transparent citations for all factual claims

Dynamic Expert Creation The network grows with every unique question:

  • Students ask about niche topics → New specialized expert created
  • Expert joins permanent network with full capabilities
  • Future students benefit from expanded expertise
  • Network continuously evolves to cover infinite subjects

The Expert Ecosystem

Our 10,000+ expert network spans:

  • Academic Subjects: Mathematics, Sciences, Humanities, Languages
  • Professional Skills: Programming, Business, Design, Data Science
  • Creative Arts: Music, Writing, Visual Arts
  • Technical Domains: AI/ML, Web3, Quantum Computing, System Design
  • Life Skills: Finance, Communication, Career Development
  • Emerging Fields: Any topic you can imagine

Each expert specializes in specific sub-domains, ensuring deep expertise rather than shallow generalization.

Installation/Run Instructions

Ensure the environment variables are set in the respective .env files of frontend and backend.

Backend (FastAPI)

cd backend
uv sync
uv run backend

Frontend (Next.js)

cd frontend
npm install
npm run dev

Tech Stack

Agent Framework: OpenAI Agents SDK - Powers our multi-agent orchestration and expert network LLMs:

  • GPT-4.1 - Expert orchestration, confusion detection, and adaptive teaching logic
  • Gemini 3.5 - Image generation and visual diagram creation Backend: Python FastAPI - High-performance agent communication Frontend: Next.js, Tailwind CSS - Responsive, modern interface Database: Neon Postgres (Serverless) - Expert profiles, student progress, and learning analytics Knowledge Base: Advanced RAG System - Domain-specific accuracy with verified sources Search Integration: Real-time web search for current information

Key Features

Confusion Elimination Engine

  • Sentiment analysis detects confusion in real-time
  • Response depth evaluation ensures understanding
  • Automatic teaching strategy pivoting when needed
  • Progress tracking across learning sessions

Personalized Learning Pathways

  • Each expert remembers your progress
  • Builds on previous knowledge
  • Adapts difficulty progressively
  • Custom-tailored explanations based on your learning style

Multi-Expert Consultation

  • Consult multiple experts on the same topic
  • Compare different teaching approaches
  • Find the explanation style that works for YOU
  • Learn from diverse perspectives

Persistent Memory

  • Student progress tracked across sessions
  • Custom expert recommendations based on history
  • Learning patterns analyzed for optimization
  • Seamless continuation of educational journey

Why Hyper Personalized Tutor?

vs. Generic AI Chatbots: 10,000 specialized experts vs. one generalist vs. Traditional Tutoring: 24/7 availability, unlimited subjects, fraction of the cost vs. Video Courses: Interactive, adaptive, and personalized in real-time vs. Other EdTech: True specialization with confusion-aware teaching

Vision

We're building the world's largest network of AI educational experts - where every student has instant access to unlimited, specialized expertise that adapts to their unique learning needs. No more confusion loops. No more one-size-fits-all. Just precision learning, powered by thousands of experts who understand exactly what you need.

About

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages