As Accenture scaled to millions of predictive models, it required automation to ensure accuracy, prevent false alarms, and preserve trust. Teresa Tung, Ishmeet Grewal, and Jurgen Weichenberger explain how Accenture implemented a DevOps process for analytical models that's akin to software development—guaranteeing analytics modeling at scale and even in noncloud environments at the edge.
This talk was originally given at Strata 2017 Singapore.
- Title: DevOps for models: How Accenture managed millions of models in production—and at the edge
- Release date: April 2018
- Publisher(s): O'Reilly Media, Inc.
- ISBN: 9781492037385