Overview

SOPHI AI is a prototype app designed to help users find up-to-date volunteering opportunities based on their interests and location. Unlike traditional volunteer matching sites, which often have outdated listings, SOPHI AI leverages Metaphor API and Generative AI to retrieve and summarize the most relevant, real-time opportunities from across the web and social media.

By integrating neural search with LLM agents, the system extends beyond the limitations of traditional keyword searches and retrieval-augmented generation (RAG), offering a more flexible and accurate discovery process. I developed the app using Python, Streamlit, and Metaphor’s free-tier API.

Key Features

• AI-driven search – Finds relevant volunteering events using neural search and LLM agents

• Real-time updates – Overcomes traditional LLM training cutoffs by pulling live web and social media content

• Flexible query design – Uses Metaphor SDK for fine-tuned search results, rather than keyword-based or rigid RAG indexing

• Event summarization – Analyzes retrieved events and extracts key details like date, location, and organization