Register
GenAI can ingest an unstructured document and answer your natural language questions about it, but it isn’t yet as good at “talking” to structured data in a database, especially if the database models complex real-world relationships. This presentation uses stories, code and demonstrations to describe how to model hierarchical data in a graph database, how to use sentence embeddings and cosign similarity to enable people to chat with the graph database, and how different LLMs deliver different levels of accuracy and completeness.
Speaker
Jonathan Lowe
Senior Director and Data Science Lead, Operations & Insights @Pfizer
Jonathan Lowe is an analytics strategist-builder with 30 years of hands-on technical and leadership experience. He envisions, sells, builds and manages fresh approaches to data-driven decision making for biopharma, agriculture, supply chain and telecom clients. He unlocks value rapidly by bridging business and technical cultures and thrives when growing new teams. He has worked at Informix, IBM, Deloitte, BCG and Pfizer, and has started two of his own companies in Berkeley, California and London, England. Sailing small boats makes him feel connected to the world.
