Case Study: Vital achieves faster, more accurate hospital wait time predictions with Chalk

Vital predicts hospital wait times with Chalk’s feature platform

Vital, a healthcare technology company, was struggling with a slow and unreliable data infrastructure that hindered its ability to update the AI models powering its patient experience software. Their previous feature platform, which relied on Spark and Databricks, made experimentation tedious and required deep specialized knowledge, limiting development to only a handful of engineers.

By deploying Chalk's feature platform within their own AWS infrastructure, Vital solved these challenges. The solution enabled rapid iteration, with the team now deploying model updates 2-4 times per month and up to nine engineers working concurrently. Key results include improved model accuracy, the ability to serve over 200 predictions per patient visit with millisecond latency, and guaranteed data privacy as no patient information leaves Vital's cloud.

Vital

Mack Delaney

Director of Machine Learning