There's a lot KAYAK is doing to stand out in Boston’s booming tech space.
Since 2004, the global company has focused on making travel better by searching thousands of other sites to give travelers as much information as possible before they book a flight, hotel or destination. They were one of Facebook’s first partners to launch a chatbot, and they’re leading the travel industry by introducing AI innovation to the equation.
Today, KAYAK funnels many of its brightest minds into AI research, hoping to build the industry’s best virtual travel agent. Since 2016, the company has hired more than 80 members of the tech team and have 18 interns starting in June.
At the forefront of KAYAK’s AI initiative is chief scientist Matthias Keller.
Keller started working at KAYAK’s Zurich office over three years ago as an analytics engineer before the company fully tapped into his vision to move KAYAK into unchartered AI and machine learning territory.
Now, Keller works in the travel company’s Cambridge office and is the unofficial “right hand” of KAYAK CTO Giorgos Zacharia.
We caught up with Keller to learn more about his role as chief scientist, KAYAK’s biggest recent wins and to hear how he envisions the travel of tomorrow.
What does your role as chief scientist entail?
KAYAK has always been a tech company that started with being able to search hundreds of sites at once using machine learning over 10 years ago. We’ve always had a team solving machine learning problems. Now, technology has changed with AI becoming more consumer facing. While we are still really laser focused on our core business of getting users the results they need for the prices they want, around one and a half years ago, we started to work on AI platforms that use voice and chat to help people plan trips. That’s what I’m working on. In the last year, I’ve focused on building our own language processing technology.
What have been some of KAYAK’s biggest projects recently?
We started doing natural language processing in December 2015 and by April 2016 we had our first public product — a slackbot that was added to the first 1,000 skills Amazon’s Alexa had. We also rolled out a “chatbot” for Facebook’s Messenger Extensions so people can search for and book travel in the app. Travelers can set it up and be notified when there’s a flight delay, or can ask what the best time to fly to Cancun would be.
Why did you want to work at KAYAK?
You never run out of interesting projects because there are so many things happening related to travel right now that we will never be done. Also, I really love the team and our culture — how we allow people to move and grow in the company. Three years ago, I was a business analyst and now I’m chief scientist, right hand to our CTO. I can’t say it happens every day but once we see a person doing work beyond his or her job’s potential, we move that person into a position where he or she can have more fun working on bigger projects. For example, our senior vice president of business development for the U.S. and Canada started as an intern on our tech team six years ago.
Wow. How would you describe KAYAK’s culture?
Fast. Everything moves quickly. Startup culture is built into our DNA as a company. We’re very data driven. When we created our Facebook chat extension, the time between knowing we would do it and having it launch was two weeks. We’re also a super fun company but at the same time super professional.
What do you look for in job candidates?
Generally we want people who are great engineers who fit into our culture. What’s even more important to us is finding people who enjoy building things and enjoy a startup-like environment. That means there’s not a form with a number for every process and we have a very flat hierarchy. There’s a lot of communication with different people. You can be brand new and you’ll know the CTO personally.
What are some future travel tech trends you anticipate?
Right now there’s no good product that can help you figure out where you should go and why you should go there on a larger scale — not which bars to visit on a trip, but why you should go to a city at a certain time. There’s too much content out there for humans to consume, so I hope more AI tech goes toward helping humans make decisions about that.
This interview has been edited for length and clarity.
Photos via company
