IntelliEarth | Devpost

HomePage
pipeline flow

Video Link

Inspiration

The need for personalized, sustainable, and equitable education inspired the development of the Sustainable Agentic Tutor. We wanted to combine technology, environmental responsibility, and educational equity.

What it does

The Sustainable Agentic Tutor is an AI-powered educational tool that:

Creates personalized learning paths using quiz responses and curriculum PDFs
Fetches real-time resources from trusted platforms like Khan Academy and YouTube
Offers multi-language support for diverse learners
Provides a live dashboard to track energy usage, carbon savings, and cost comparisons, promoting sustainability

How we built it

We used self-hosted large language models (LLMs) and a retrieval-augmented generation (RAG) pipeline to deliver dynamic, tailored content, served on a Python/FastAPI backend. The project also integrates a live dashboard that tracks the environmental impact of learning sessions. We built the frontend with Next.js and TailwindCSS, ensuring a responsive and interactive user experience.

Challenges we ran into

Optimizing the AI for sustainability:
Self-hosting LLMs required balancing performance and energy efficiency.
Integrating backend and frontend:
The complexity of the backend required multiple iterations of the design and structure of our endpoints. To ensure speed, we optimized data retrieval, minimized redundant API calls, and implemented caching where necessary.
Pivoting from a TypeScript backend to Python: The team originally planned on using a TypeScript backend, directly integrated with the frontend using Next.js API routing. Realizing the limitations of using a TypeScript backend, we decided to pivot to Python and FastAPI. This was a challenge because most of the team had little prior Python experience.
Ensuring accessibility:
Developing a system that supports multiple languages and is user-friendly for students from underprivileged backgrounds was challenging.

Accomplishments that we're proud of

Successfully created a self-hosted, eco-friendly AI-powered tutor
Implemented a real-time sustainability dashboard
Made quality education accessible to students worldwide through free, open-source resources

What we learned

We learned the importance of integrating sustainability with AI, as well as how to create personalized, engaging educational experiences for diverse learners. We also gained insights into optimizing AI models for energy efficiency.