🧠 EconoMind: Reimagining Education with AI Agents
EconoMind is an interactive educational platform designed to radically simplify and personalize economics learning. This project is my submission for the Cloud Run AI Agents Hackathon.
💡 Inspiration and Vision
The project was inspired by the belief that AI can transform education by moving beyond static theory. My goal was to leverage my expertise in economics to build a solid foundation that could eventually be expanded to other disciplines.
🏗️ How We Built It: The Dual-Agent System
We built a specialized dual-agent system using the Google Agent Development Kit (ADK) and Gemini models:
- 🎓 The Professor: An expert persona focused on providing clear, contextual explanations and theories (e.g., GDP components, Comparative Advantage).
- 💪 The Coach: An action-oriented persona that bypasses theory to generate instant, practical exercises and concrete case studies.
This architecture is designed to facilitate a complete learning loop (theory to practice) within a single platform.
🚀 Architecture and Deployment
The application is deployed as a scalable microservices architecture on Google Cloud Run:
- Frontend: React (Vite/Tailwind CSS)
- Backend: Python FastAPI API, responsible for running the ADK agents.
The core challenge was ensuring strict separation of the agents' roles—the Coach must never explain, and the Professor must never assign an exercise. We solved thGoogle Gemini, Google ADK (Agent Development Kit), Google Cloud Run, Python, FastAPI, React, Docker, Tailwind CSS, Vertex AI.is through careful prompt engineering and defining distinct agent functionalities within the ADK.
🎯 Our Long-Term Roadmap
Our vision is to evolve EconoMind into a dynamic multi-agent system (V2.0): implementing stateful conversations (memory), enabling agents to track progress, and allowing them to collaborate to build truly personalized learning paths.
For detailed technical documentation, please refer to the Source Code link below.
Log in or sign up for Devpost to join the conversation.