Amazon Machine Learning is designed to remove barriers to entry for developers, said Matt Wood, GM of the Data Science Team at Amazon Web Services, Inc. This fully managed machine learning service is specifically designed to enable developers to “focus on working with their data,” said Wood. Wood explained that Amazon Machine Learning connects developers with data they already have through experiments using “visualization and interpretation tools.” Once they’ve gained a more comprehensive understanding of their data, Wood explained that developers can then begin to bring predictive models.

History of data in organizations

Three to five years ago “the role of data was very different inside organizations,” Wood said. Data was being generated at sufficient scale, but at that time, organizations had trouble managing it.

The trouble lay in their data centers. The problem was that the “walls of those data centers couldn’t’ move” and the their “resources couldn’t change,” Wood stated. Essentially, Wood said, they were “frozen in time,” bound by constraints that made them unable to experiment with their data or ask it the questions they needed to.

Cloud was the catalyst that helped “move those constraints.” It allowed organizations to be “more productive with their data.”

Machine learning as a service will change the organizations interact with data

As previous year’s constraints melt away, Wood explained that “more developers are interacting with data on a day to day basis.” He laid out three ways that developers have begun to use data:

1. Using data warehousing to look retrospectively at what’s happened on their platform. This includes tasks like log analysis and using Splunk to analyze operations.
2.Using real time data streaming, like Kenesis, to look at the data. Tools like dashboards let developers see what’s happening right now in their organizations.
3. Using machine learning to identify patterns in large amounts of data. Those patterns help developers figure out what’s going to happen next.

While Amazon Machine Learning is designed to make life easier for developers, Wood remarked that data scientist also benefit. It allows them to “add more of their differentiated skills and remove some of the undifferentiated heavy lifting,” Wood said. Specifically, Wood said it’s faster to build summary stats and interpret, instead of develop the cluster.

How Amazon MLaaS works

Machine learning as a service is a simple process, Wood explained: “customers supply the data” and then the product takes it from there. “The platform,” he detailed, “will go of and it will learn the structure of the data for you.” Then, it will “start to make recommendations on transformations it can do to the data to make it more applicable,” according to Wood. Customers then validate the model before putting it into production.

Machine learning acts like a flywheel

Machine learning is so valuable because “customers value iterations [and] being to abel fail quickly.” When customers are able to experiment in innovate ways, it’s natural that a few experiments should fail. Getting to “the one that works as quickly as possible is extremely important,” stressed Wood.

Success breeds more success, he observed, saying “It’s a fly wheel.” With more algorithm development, there will be “more places to use it.” In turn, those different applications feed off each other and make it easier to “break down the biggest barriers to adoption of machine learning by making a service that is high scale, high performance, but is beautifully easy to use,” Wood said.

Watch the full interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of AWS Summit 2015.