Skip to content

nikisetti01/ADK-edu-suite-for-gdg-Milan-hack-2025

Repository files navigation

🎓 ADK Educational Multi-Agent Suite

A modular backend suite demonstrating how to use agents built with ADK (Agent Development Kit) and tool calling to create interactive and intelligent educational systems.

Each module shows a real-world use case where multiple agents collaborate via structured communication protocols to support and enhance the student learning process.


📁 Project Structure

Folder Description
calendar/ Creates a personalized study calendar for an exam or goal using a planning agent that organizes tasks and dates.
mock_exam_generator_app/ An agent that generates mock exams from examples provided by the student.
question_agent/ Helper agent for question formulation and analysis.
socratic_agent/ A multi-agent system that engages the student in a socratic dialogue loop, refining responses through targeted questioning and performance analysis.
tester_agent_app/ An agent that grades exams and provides personalized feedback on what to review.
.env.example Sample environment configuration file.
requirements.txt Python dependencies needed to run the project.

🧠 How It Works

Each module consists of:

  • A FastAPI application (fastapi_app.py) exposing agent functionality through a REST API.
  • One or more ADK agents that collaborate to complete educational tasks (e.g., correction, planning, tutoring).

Agents communicate using a message-passing protocol, enabling scalable and coordinated multi-agent workflows tailored to education.


🚀 Highlighted Examples

🧠 socratic_agent/

Implements a full tutoring loop:

  1. A question agent asks a preparatory question.
  2. The student responds.
  3. An analysis agent identifies weak points.
  4. A socratic agent generates a new question focused on the gaps.
  5. A fitness agent evaluates the learning progress:
    • If improvement stagnates → the student is redirected to study relevant material.
    • If fitness improves → the student receives a new, unrelated question.

The fitness agent also uses a penalty function to gamify progress and maintain challenge.


📅 calendar/

Builds a personalized study schedule based on:

  • Student availability
  • Goal deadline (e.g., exam date)
  • Current preparation state

⚙️ Setup Instructions

Clone the repository:

git clone https://github.com/your-username/adk-edu-agents-example.git
cd adk-edu-agents-example

About

Educational multi-agent apps using ADK and FastAPI — built at GDG Hackathon Milan 2025

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •  

Languages