How one data scientist’s side project led to a once-in-a-lifetime job

 

Jordan Meyer, whose three-person team of data scientists bested more than 3,800 teams from 91 countries in January to win the $1 million Zillow Prize competition to improve the Zestimate, is on to his next challenge: a full-time role at Zillow advancing the algorithms behind the company’s new business of buying and selling homes.

The path to his new role wasn’t exactly a straight line. It started off with a bang – or more precisely, a surprise – when Jordan learned he and his team had won Zillow Prize. The grand prize win capped nearly two years of competition in one of the most popular machine learning contests ever on Kaggle, the platform that administered the competition.

“Everyone knows the Zestimate, whether you’re a data scientist or not,” Jordan said at the time of his team’s win. “It’s still unbelievable that we won the million-dollar prize. I feel so lucky and so proud of all our hard work.”

At that point, Jordan didn’t know where that hard work would lead him. Working remotely as a data science consultant from his home outside of Raleigh, North Carolina, he advised companies on how to apply artificial intelligence to address their business challenges, improve operations, and quickly and efficiently build predictive machine learning models. He’d spent most of his career in analytics jobs and had a passion for computer science, working with data, and using technology to solve problems. 

Even his hobbies applied his passion for machine learning, like building an algorithm to discover unique spirit and cocktail combinations that he used to improve his home bar collection. He was also a long-time member of the Kaggle data science community, where he loved learning from expert engineers and data scientists and seeing the diverse set of problems Kaggle competitions aim to solve. Though he’d never participated in a competition before, when Zillow Prize launched in May 2017, the potential of winning $1 million to improve the Zestimate convinced him to enter the initial qualifying round.

As he progressed through the competition and spent more time working on the Zestimate outside of his day job, the more he saw the depth and potential of machine learning and data engineering to solve a single, complex challenge.  

“I’d always consulted, which meant switching industries to fit different customers’ needs,” he recalled. “The only thing that was constant in all of my work was math. Zillow Prize was the first time I ever spent over a year on any one problem.”

Jordan estimates he spent 400 hours on his contributions to the team’s winning algorithm. But those hours spent developing a deep knowledge about real estate data and practicing valuing homes with AI gave him an unparalleled advantage when an opportunity opened on the machine learning team for Zillow Offers, Zillow’s home-buying and selling program.

“Winning the Zillow Prize was definitely a surreal experience,” says Jordan. “But I’d say my life changed the most with this job.”

“We just couldn’t pass up hiring Jordan,” said Dr. Stan Humphries, Zillow Group’s chief analytics officer and creator of the Zestimate. “He and his Zillow Prize teammates impressed us with their creative ideas and nuanced machine learning techniques. I got to meet Jordan when I let him know that he’d won the Zillow Prize, and I knew then that we needed to hire him. We’re excited for him to bring the same creativity and technical skills to our artificial intelligence and machine learning work, which is at the heart of our efforts to re-engineer and simplify how people buy and sell homes.”  

As Jordan settles into his role, he already sees how his current work presents similar challenges he encountered during Zillow Prize.

“Instead of figuring out how to create the best estimate on every home like the Zestimate does, we’re working to come up with new and creative ways to accurately price individual homes – ones that we have a bit more data directly from the seller when they request a Zillow Offer,” he shares.

Through Zillow Offers, homeowners in eligible markets can request a no-obligation cash offer on their home from Zillow, helping remove the stress and uncertainty of selling a home. To help power that simplified real estate experience, Jordan’s working to build an AI-based system of neural networks capable of learning about millions of homes across the country, which then uses those learnings as background knowledge when pricing individual homes.  

“For most people, their home is the most significant investment of their lives,” he said. “Using AI to make the home-buying process easier is a great opportunity and responsibility.” 

Family-friendly benefits land Zillow Group on Fortune’s Best Workplace for Parents List