Engineering Spotlight: FeatureX
FeatureX is using machine learning and other AI technologies to analyze global data sets. With their sophisticated technology, FeatureX is able to discover and predict patterns in satellite and economic data.
We connected with FeatureX’s founder and CTO, Gil Syswerda, to talk about the engineering team and how they are able to take on a job of this magnitude. We even had a chance to chat about how the company built the engineering team and what the company’s internal culture is like.
Also, FeatureX is hiring! Check out their BIZZpage for all of the company’s openings!
Quick Hit Details
Year Founded: 2015
Number of employees: 10
Number of engineers: 7
Industry: machine learning technology
FeatureX is a cool name. Is there a meaning behind it?
In machine learning, success begins and ends with creating the right features, which are individual measurable characteristics of an event or dataset, and X marks the spot!
Can you share the details on what FeatureX does?
We analyze large data sets using machine learning, extracting economic activity over time. The primary dataset we tackle right now is millions of high-resolution satellite images, where we take a sequence of satellite images and extract a sequence of economic activity. We look for patterns over time and these patterns are the basis for insights into human and natural activity, as well as creating economic prediction models.
Your website talks about using machine learning to discover meaningful patterns in the world’s data. Can you share what that means?
Machine learning, combined with the massive amounts of affordable computing power made available via cloud computing services and GPUs, make it possible to find patterns in very large datasets. Also, by using our proprietary machine learning technologies, we can see patterns in data that no one else can see.
What are some of the different technologies that the engineering team gets to work with and at what scale?
We make extensive use of deep learning, including autoencoders, GAN, and CNNs. We are very innovative in that space, and have developed unique architectures and operators. We also employ a proprietary machine learning system, unlike any publicly available system, that is very effective at finding certain kinds of patterns in data.
What are some of the interesting projects that the engineering team is tackling?
In the satellite and remote sensing space, we are building a set of platforms that will allow both ourselves and others to create SaaS products easily and effectively. These products will provide business insights for financial firms, insurance companies, large corporations and governments. Our goal is to empower customers to make globally-informed business decisions by allowing them to easily and inexpensively acquire and analyze satellite images and other remote sensing data. Clients can build web-based applications for either internal use or for the development, sales, and deployment of SaaS applications worldwide. These organizations can focus on their business needs while we handle the licensing and acquisition of images and do the heavy lifting with regard to image analysis and computer vision.
Can you describe the type of background or experience that you generally look for when hiring for the engineering team?
At our heart, we are a machine learning research and development team. As such, we look for research scientists who can innovate in either the machine learning or computer vision areas. We also look for research engineers who love to code, but also have an interest in machine learning and want to learn more.
On the platform side, we look for people that love software development and who think naturally in terms of service oriented architectures (SOA) and APIs.
We hire all experience ranges, from people with deep industry experience to people just graduating from school. We value experience, but above all we value being bright and innovative.
How would you best describe the culture and working environment for the engineering team at FeatureX?
We are very focused on learning. We are working at the cutting edges of both machine learning and SaaS software deployment, and in both worlds the pace of innovation in the outside world is high. We are committed as an organization to staying current. As a result, we have a collegiate environment where ideas are shared and discussed, and a lot of cross-team dialog happens. We host our own machine learning seminar each week, and invite selected outsiders to participate. We read a lot of papers and attend scientific and technical conferences. One of our new employees, recently out of school, commented to me recently that he thought the pace of learning in grad school was high, but that grad school didn’t compare to FeatureX!
In terms of day to day life, we have great offices in Kendall Square where you can often find our employees hanging out, playing foosball or board games. That’s another important part of our culture here. Everyone respects and values the contributions of their teammates, which makes for a collaborative work environment, but also a lot of fun. We do company-wide events like bowling or kayaking, but the inherent friendliness and interest in spending time with each other is what makes it a truly enjoyable place to work.
What can someone expect during the interview process?
As a candidate, you start by having an initial phone conversation, usually with me. In that conversation, we establish that there is mutual interest and we get to know one another a bit. That is followed by a programming test, where we give you a small project to code up. In the coding test, we test your ability to understand a problem statement, and to devise a reasonable programming solution to that problem. We then do a technical video call lasting about an hour. In that interview, usually with 3 people on our end, we do an assessment of the skills and knowledge you bring to the table. Depending on what position you are applying for, you will be asked questions about computer science, software development, math, statistics, machine learning, and computer vision. The next step after that is an onsite visit, where you will spend a good part of a day with the team. During this visit, we will delve into more technical questions and also determine, from your perspective and ours, whether we are a good cultural fit for each other. You are encouraged to meet one on one with our employees to get an inside perspective on what it is like to work at FeatureX.
What does your engineering team do for fun outside of work? Are they involved in any local tech organizations?
We are a pretty social group. A number of us work out together in the mornings (we provide free gym membership!). We often eat lunch together, and have Friday evening gatherings with food and drinks. The Thursday afternoon machine learning seminar has become an important part of our culture as well, and we have regular outside people who attend. In some sense, we ourselves have become a tech organization that others seek to be a part of!
We are located in Kendall Square on purpose. We want to be next to MIT and be part of that community. We are affiliated with CSAIL, we sponsor the ML Tea, and we attend talks and events all the time.
As for non-work related fun? A lot of our people like to be outside - hiking, biking, running. We have a few who enjoy cooking and travel as well. There’s undoubtedly a fair amount of video game interest too.
Are there any other random fun facts that you’d like to share?
We’ll use practically any excuse to have a party. Early on, we had an issue with some financial data that we were using to supplement our models. We called it the 2006 problem. When we finally figured it out, Emily and Colleen, who make up our HR team, threw a surprise party, complete with a 2006-themed cake and champagne!
Birthday parties, Chinese New Year, 4th of July (you could see the firework from our office), parties to celebrate personal or work achievements - nothing is off limits!
Rapid Fire Q&A
What’s on tap?
Changes by the week! There’s always an IPA, Guinness, and cider, but you’ll see anything from specialty cocktails to seasonals - we try to get creative.
Star Wars or Star Trek?
We have employees from all over the world. It’s not so much Star Wars or Star Trek as it is “zee” or “zed”, “chips” or “crisps”?
iPhone or Android?
Coffee - hot or iced?
Either - we’ve figured out how to use the fancy espresso machine in the kitchen and the world is our oyster (or roaster, as the case may be).
Favorite employee perk?
Free gym membership is a huge hit. But traveling worldwide for conferences isn’t bad either.
What TV show describes the engineering team’s culture?
There might be varying opinions on it, but every single one of us has watched Westworld. I’ll just say that.
“I really like the research environment at FeatureX. It’s stimulating! I have to tackle a lot of very interesting or challenging problems that use machine learning in a novel way and I know that I'm doing important work in leveraging satellite images. Doing this alongside a really nice cohort of people makes the whole process a lot more fun. I have a lot of ownership over the problems I'm solving but also a lot of collaboration and opportunity to learn new things from others.”
Vision Research Scientist
"I joined FeatureX because I was looking to challenge both my strengths and weaknesses. Since joining, I have had the ability to nourish my coding expertise, while at the same time learning all about statistics and machine learning. Through the many internal machine learning discussions that take place as well as the weekly seminars (where we present and dissect an industry paper) I have had the opportunity to ramp up quickly in the machine learning space."
Senior Software Engineer
“I’m really enjoying my internship here. First of all, the employees at FeatureX really know how to do research, and it’s not just bottom-line focused or narrowed into churning out product that may or may not work. We’re meticulous about how we take the steps, and Gil is really focused on that which is great.”
Machine Learning Intern
Images courtesy of FeatureX.