Tecton Offers Its ML Platform Through Google Cloud Partnership

The Tecton-Google Cloud combination will help solution providers accelerate the data development work associated with customers’ machine learning projects.


Machine learning startup Tecton has struck a strategic partnership with Google Cloud under which the Tecton Feature Platform, used to supply machine learning models with high-quality data, is available for Google Cloud users.

Through the partnership the Tecton Feature Platform is available to work with Google Cloud’s AI and data services to speed up the development of machine learning models while controlling costs, according to Tecton.

Tecton is now available through the Google Cloud Marketplace, meaning customers can use Google Cloud credits to purchase the service. (Tecton has a similar partnership with Amazon Web Services.)

Sponsored post

[Related: Machine Learning Startup Tecton Raises $100 Million In LatestMachine Learning Startup Tecton Raises $100 Million In Latest Funding Round Funding Round ]

“This kind of technology wasn’t really available for Google Cloud customers,” said Tecton co-founder and CEO Mike Del Balso in an exclusive interview with CRN. “And now we’re bringing that solution to Google Cloud customers for the first time.”

Tecton, based in San Francisco, was founded in 2019 by the developers who created Uber’s Michelangelo machine learning platform. The company has raised $160 million through multiple funding rounds.

Tecton’s platform is used to automate the process of collecting, preparing, managing and updating the huge volumes of high-quality data needed to train machine learning models and provide them with data when they are in production for real-time predictive and generative AI applications – everything from pricing applications, customer scoring and recommendation engines to automated loan processing and fraud detection systems.

“These are just common types of decisions that a lot of businesses have to do. And they have to do them at scale. And they have to do them really quickly. And they have to do them very reliably,” Del Balso said.

Developing the pipelines to transform batch, streaming and real-time data into the data signals – commonly called machine learning “features” – that machine learning systems need is a complex process and a reason why machine learning projects often fail.

“It’s not uncommon for our customers to have literally thousands of features that power their models,” Del Balso said.

Google Cloud provides its Vertex AI system for training and deploying machine learning models and customizing large language models for use in AI-powered applications. (It also offers open-source tools such as TensorFlow and Kubernetes.) Google Cloud’s data processing infrastructure services, including DataProc and BigQuery, also come into play within many ML projects.

The Tecton platform can be used as the “connective fabric” to tie these systems together for building production-ready ML features, according to Del Balso, automating the ML features lifecycle through feature definition and data transformation to online serving and operational monitoring.

Using the Tecton platform helps developers build better machine learning models by using better data, Del Balso said. By automating the data transformation and management stages of machine learning projects, they can put ML systems into production much faster. And the Tecton platform provides a layer of enterprise management and collaboration that’s often missing within ML initiatives, the CEO said.

Solution providers and strategic service providers who perform AI and machine learning development work for customers can use the Tecton-Google Cloud combination to do that work more efficiently. “This is just another option to make their customers more successful,” Del Balso said.

“We are delighted to partner with Tecton to bring advanced Machine Learning feature engineering capabilities to Google Cloud,” said Manvinder Singh, Google Cloud managing director, partnerships, in a statement. “Via this partnership, customers can further accelerate the building of machine learning applications via Tecton and Google Cloud’s AI and data services.”