Your credit card company can almost instantly detect suspicious transactions. Your child’s Barbie talks back to her like a real friend. UberEats is able to accurately predict when it will deliver your takeout.
In today’s data-driven world, technological advances like these have become increasingly common. Computers are getting better at analyzing data on the fly and using that analysis to make predictions — thanks to specialized software engineers who are trained in machine learning.
Machine learning engineers use artificial intelligence techniques to design complex prediction models. Their work allows computers to detect patterns in data and use that information for predictive purposes, such as product recommendations or speech recognition. In the credit card scenario, a machine learning algorithm uses cardholder data to determine what typical spending patterns are and sends a warning or blocks the transaction when something diverges from that pattern.
“In conventional software development, you’re writing an algorithm — telling the computer step by step what to do,” explained Michael Friedman, lead software engineer at Salesforce. “In machine learning, you give the computer some kind of machine-learning technique, and it looks at the data and learns how to analyze the data by itself.”
jobs in machine learning on the rise
The U.S. Bureau of Labor Statistics reports that jobs for computer and information scientists are continuing to grow apace, with employment expected to increase by 19 percent by 2026. And those with AI skills are particularly in demand: Indeed reported that AI-related job postings leapt by 119 percent between 2015 and 2018.
Engineers with expertise in machine learning could keep a career edge for at least the next 10 years, according to Dave DeBarr, principal applied researcher at Microsoft.
“We’re talking about automation and improved customer experience across many different domains — everything from medicine to agriculture,” DeBarr said.
DeBarr and Friedman, both instructors for the Certificate in Machine Learning offered by UW Professional & Continuing Education, said graduates are prepared to be high-value employees for any company that’s applying data science to solve business problems.
19% expected growth from 2016-2026 (faster than average for all jobs)*
Median Annual salary
Source: Bureau of Labor Statistics
There’s ongoing proof of this hot job market among Seattle area employers. Companies like Amazon, Microsoft and Facebook are all hiring machine learning engineers. These professionals are also in demand at Zillow, which in 2017 launched a contest to improve its popular real-estate value algorithm. The prize? More than $1 million.
what machine learning engineers need to know
DeBarr said machine learning engineers need to know about five areas of data science: data collection, data management, metrics, modeling and experimentation.
“Machine learning is the modeling,” DeBarr said. “The most important knowledge is how to try out a number of models for a particular task and pick the best one.”
He said the Certificate in Machine Learning also covers state-of-the-art techniques in neural networks and deep learning, a subfield of artificial intelligence inspired by the workings of the human brain.
If you already work as a statistician or a software engineer — or if you’re entering the field with background in math and programming — DeBarr said you’ll be in a good position to learn more about the calculus, linear algebra, probability and statistics concepts you’ll need to work in machine learning.
“With the internet, we’re collecting so much data about users,” Friedman said. “The idea of machine learning is that you can somehow interpret that data and use it to give the user a better experience. In the evolution of software development, it’s the next stage.”
Interested in how expertise in machine learning can advance your career? Visit the Certificate in Machine Learning web page to learn more about the program.