Machine Learning is used in several ways at Regions – from detecting and preventing fraud, to providing customer insights and reducing risk.
As Machine Learning has grown at Regions, so has the need to scale and operationalize machine learning models.
At Regions, Daniel Stahl, head of Data and Analytic Platforms, is responsible for leading Machine Learning Operations.
Stahl was recently a featured speaker at the Robust & Responsible AI Summit presented by WhyLabs, an AI observability platform that provides organizations with monitoring of machine learning models. His presentation, entitled “Operationalizing Machine Learning at a Large Financial Institution,” addressed how Regions is building a scalable, replicable model for Machine Learning Operations.
“One of the industry’s struggles around machine learning and AI is how to scale development and deployment,” explains Stahl. “An oft-quoted statistic is that 80 percent of machine learning projects never make it into production. Regions uses a combination of culture, approaches and tools to create a ‘factory’ for AI development and deployment.”
Stahl believes that this type of environment enables the efficient and scalable operationalization of machine learning models that far exceeds industry norms.
3 Keys to a Successful Machine Learning Operations Environment
So how does Regions create this environment? In his presentation, Stahl breaks down the three keys to a successful machine learning operations environment:
Part of this discipline includes practices that encourage teamwork and collaboration.
“One of Region’s core values is to enjoy life – have fun – and I think we really embody that. We’re very collegial and collaborative and have a can-do attitude where we are always
looking for innovative ways to solve problems and do it in a way that brings everyone along for the ride.”
Watch Stahl’s presentation below to learn more about Machine Learning Operations at Regions.
Related data and analytics stories on Doing More Today: