Session
Building AI That Matters: Powering Global Healthcare at The Virtue Foundation
Overview
| Experience | In Person |
|---|---|
| Track | Artificial Intelligence & Agents |
| Industry | Healthcare & Life Sciences |
| Technologies | Lakeflow, Unity Catalog |
| Skill Level | Beginner |
Ever wonder how The Virtue Foundation connects medical resources with underserved regions? As the first collaboration under Databricks for Good, the foundation partnered with us to build an AI-powered platform improving medical outcomes across 73 countries. This talk breaks down the production Lakehouse architecture behind the system that continuously curates healthcare data to match volunteers with communities in need.What we’ll cover:LLM extraction at scale: Processing 5M+ web pages with OpenAI models via bespoke multimodal pipelines running natively on Spark, reducing cost while improving precision.Data engineering on Spark: Running skewed, multi-terabyte workloads using Spark for parallelism, Photon for performance at scale, and production-grade orchestration.Making data trustworthy: Unifying millions of healthcare facilities through entity resolution.A candid, systems-level look at what it takes to move LLM-powered data extraction from proof of concept to global impact.
Session Speakers
Nicolas Douard
/Tech Lead
Virtue Foundation
Priyanka Mehta
/AI FDE
Databricks
Michael Berk
/Staff AI FDE
Databricks