Olena Shevchenko on Experimenting with Neo4j to Build a Personal Knowledge Agent

Voices is a new mini-series from Humanitarian AI Today. In daily five-to-fifteen minute flashpods we pass the mic to humanitarian experts and technology pioneers, to hear about new projects, events, and perspectives on topics of importance to the humanitarian community. In this flashpod, Olena Shevchenko, a data scientist and machine learning engineer, speaks with Humanitarian AI Today producer, Brent Phillips, about Wintertime conditions in Ukraine and her experimentation with Neo4j, a graph intelligence platform, to build a "Personal Knowledge Agent.” Seeking a more structured way to retain and reuse knowledge, Olena explains how her application converts unstructured text into a network of "concepts" and "connections," allowing for precise, context-aware answers to user queries. Her project highlights uses of Neo4j combined with Local Large Language Models (LLMs) and Retrieval Augmented Generation (RAG). Olena and Brent help to provide staff from Humanitarian organizations with an introduction to Neo4j and how knowledge graphs could help transform humanitarian aid by mapping complex relationships between organizations, people, and activities across global crises.