AT&T Deploys AI Feature Store
November 21, 2022
updated its AI Feature Store, an integrated online computational and storage
platform that gives AT&T’s data scientists a unified, reliable source of
reusable features which are the building blocks for machine learning (ML) and
artificial intelligence (AI).
AT&T and H2O.ai co-created and deployed the platform to AT&T’s Chief Data Office (CDO) and then across AT&T last year. AT&T estimates AI Feature Store has improved speed to market by as much as 70%.
”Feature Store is an ideal platform for making our AI capabilities reusable and easy to access for our data scientists and citizen developers alike, and has helped integrate AI into the fabric of our business,” said Andy Markus, chief data officer at AT&T.
The fully managed AI Feature Store platform enables users to register, store, update, discover, use and share within AT&T highly curated, validated and model-ready historical and real-time features (attributes derived from raw data that predict an outcome such as churn, likelihood to buy, predicted demand, etc.) collected from throughout AT&T.
Conventional wisdom indicates data scientists and machine learning (ML) experts spend about 80% of their time generating, preparing and labeling data. That leaves only 20% of their time for the important tasks of analyzing and interacting with the data to solve real-world business problems. AI Feature Store flips that breakdown of time.
In the first year of launch, the co-developed AI Feature Store addressed the complexity of data quality requirements by providing a central repository that connects information across disparate systems to bring all of an organization’s data assets together in one place. AI Feature Store plays a key role in AI governance by ensuring that the right kind of data is only accessible to the right folks and taking care of sensitive data like PII or SPI with encryption, data masking and other technologies.
“The external version we co-developed with H2O.ai can help other companies embarking on their own AI transformations, even as the speed and complexity of the challenges they face accelerate,” said Markus. The acceleration of speed to market gives AT&T some of the most advanced AI and CDO departments in the entire country. The team was awarded a top spot by H2O.ai during the conference, which noted their strong AI program.
“AI Feature Store is another way to fuel our drive to help large enterprises transform themselves,” said Sri Ambati, CEO and founder of H2O.ai. “Our drive to enable transformation is at the heart of our mission at H2O.ai– we’re here to connect the world and democratize AI across organizations so that it can have a positive impact on society.”