Eco connect

Inspiration

The project was inspired by the need for intelligent collaboration in environmental sustainability. While many professionals work towards climate solutions, finding the right collaborators remains a challenge. We envisioned an AI-powered system that could strategically connect changemakers based on expertise, interests, and impact potential.

What We Learned

Through this journey, we gained deep insights into Retrieval-Augmented Generation (RAG), semantic search using FAISS, and MBTI-based personality analysis. We explored LLM-driven strategic matchmaking, graph-based collaboration visualization, and real-time data processing with Firebase.

How We Built It

We used FAISS for vector indexing, MiniLM & Sentence Transformers for embedding generation, and Gemini LLM for intelligent recommendations. The frontend was built with Streamlit, while PyVis enabled dynamic network visualizations. We also trained Logistic Regression (78% accuracy) for MBTI prediction to enhance compatibility scoring.

Challenges Faced

Optimizing vector search for scalability and efficiency.

Fine-tuning MBTI predictions to improve accuracy.

Integrating real-time insights while maintaining performance.

Ensuring intuitive UI/UX for a seamless collaboration experience.

This project is a step towards AI-driven, impactful networking for global environmental initiatives. 🚀