Expansion into enterprise markets drives AWS re:Invent machine learning debuts

Expansion into enterprise markets drives AWS re:Invent machine learning debuts

When Amazon Web Services Inc. held its re:Invent conference one year ago, Chief Executive Andy Jassy realized afterward that he had spent 75 minutes just on the machine learning section alone.

That prompted AWS to devote an entire keynote to the subject during today’s re:Invent conference, and the result was a set of announcements that reinforced AWS’ significant commitment to the machine intelligence space.

“AWS is really trying to simplify machine learning and use machine intelligence in all applications,” said Dave Vellante, co-host of theCUBE, SiliconANGLE Media’s livestreaming studio. “They’re trying to simplify the data pipeline. They’re expanding data access to the line of business, and that’s a key point.”

Vellante spoke with John Furrier, co-host of theCUBE, during the opening keynote analysis on the second week of AWS re:Invent. They discussed the company’s strategy for machine learning expansion, developer adoption of its cloud-based tools, newly announced partnerships with key industry players, edge computing solutions and AWS’ continued pursuit of new markets. (* Disclosure below.)

Targeting the developer

A central theme of today’s announcements, led by Swami Sivasubramanian, vice president of Amazon machine learning, was the company’s desire to extend artificial intelligence and machine learning deeper into organizations. Much of the keynote’s focus was on new tools for preparing and training machine learning models.

“They are really targeting the developer,” Furrier said. “There’s more horsepower required to process training and modeling, and this will lead to expansion of the personas who are going to be using machine learning.”

Sivasubramanian offered statistics that showed AWS was already well along in capturing the hearts and minds of developers. Two popular machine learning tools — TensorFlow and PyTorch — had cloud-based run rates of 90%-plus on AWS.

“The adoption of PyTorch is telling,” Furrier said. “The fact that you have cloud-based 92% adoption for TensorFlow and 91% adoption for PyTorch on AWS is a huge validation.”

Among the announcements made today was the launch of Amazon SageMaker Data Wrangler as an easier way for developers to prepare data for machine learning. Of perhaps greater significance were the three companies that will integrate services into the new AWS offering. Snowflake Inc., Databricks Inc. and MongoDB Atlas will all become integrated into the AWS SageMaker Data Wrangler tool.

“These are three platforms doing very well in the market and holding a lot of data,” Vellante noted. “They’re really streamlining a lot of specialized activities, abstracting away IT complexity.”

Focus on the edge

This week’s announcements from AWS around machine learning were also clearly targeting the edge computing sphere. The company’s press release, which briefly summarized Tuesday’s Amazon SageMaker news, mentioned “edge” 25 times.

The latest releases include Amazon SageMaker Edge Manager, which will deliver machine learning model monitoring and management for edge devices. This latest tool will undoubtedly appeal to developers as well.

“They talked a lot about doing AI inferencing at the edge,” Vellante said. “I really like Amazon’s edge strategy. The developers are going to be the linchpin to the edge.”

By extending machine learning to more builders in IT organizations, AWS is also advancing deeper into the enterprise itself. This may drive vertical expansion as well, as seen in the firm’s expansion of its Amazon Lookout tool today to now include sensors on field equipment to generate alerts.

“It’s where we’re starting to see the land grab; it’s Amazon going after infrastructure as a service and platform as a service,” Furrier said. “Clearly every vertical is up for grabs.”

Even professional football has not gone untouched by the reach of AWS. The company has a longstanding relationship with the National Football League, providing its “Next Gen Stats” platform and working with the organization to use machine learning and improve player health and safety. A top NFL executive was among the keynote presenters today.

“This is a meaningful partnership with the NFL,” Furrier said. “It is a tell sign that Amazon is continuing to lead. You’ve got the tech, why not use it? And they are.”

Here’s the complete analysis, part of SiliconANGLE’s and theCUBE’s coverage of AWS re:Invent. (* Disclosure: Amazon Web Services sponsored this segment of theCUBE. Neither AWS nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: AWS/livestream