 Hi, welcome back to theCUBE here at Stratoconference. I'm Jeff Kelley from Wookie Bun, filling in for Dave Vellante and John Furrier, who are off taking a much needed lunch break at the moment. I'm joined by a very cool guest, Jeremy Howard from Kaggle. Welcome to theCUBE, the CUBE alum, been on before. I'm delaying my lunch break to be here. Well, we appreciate it. So that's how much I like you guys. We appreciate it, thanks so much. So why don't we start off, tell people that are maybe not familiar with you, what Kaggle's all about. It's a really interesting story. Sure, I mean, maybe a lot of people listening might know about the Netflix prize, was a $1 million challenge. About 50,000 people entered it to try and improve the Netflix recommendation system. Challenge was successfully run and Netflix got a lot of value out of that. It was a very popular project. I think it's in the New York Times about six times. It really captured the imagination of a lot of people. And we realized that this kind of value was actually a great way to do predictive modeling everywhere, in science, in industry, in government. So we set up a site that allows anybody to run their own predictive modeling competition, their own Netflix prize, if you like, but rather than needing a year with a team of expert statisticians to set it all up, you just fill out like a wizard, a five-step wizard. And at the end of that, you've got yourself a predictive modeling competition. Wow, so it's kind of a self-service model. It is a self-service model. There is a team of data scientists on staff who can help with stuff like, so for example, we did a big project for NASA, which was trying to help them solve a problem that's been around since 1934, which is mapping the dark matter in the universe, that matter which seems to be there scientifically, but we can't see. They combine their domain expertise with our data science teams' skills to find a way of kind of putting that into a form that it could be run as a machine learning competition. And then they ran that on the site, went to our 30,000 PhD level data scientists, and they crunched on the data and improved NASA's best model by 300%. Wow, cool. So talk a little bit about kind of the need you're filling. We hear a lot about the skills gap. There's just not enough data scientists to go around. So a lot of your clients come to you because they don't have the internal expertise. Either they can't find them, they can't afford to put them on staff. What's the typical client? Sometimes that's true. I mean, but if we look at a number of clients, so for example, NASA on that team, at least to the advisory team, was the people that wrote the books that I learned stats from. I mean, they have people. Another of our clients, very, very innovative, very, very successful client is all state insurance. They've got a team of 30 or 40 of the world's best actuaries. Another of our clients, a little startup called Jetpack, is actually one of the co-founders, is a guy who literally wrote an O'Reilly book on data science and write software for them. In all these cases, they still come to Kaggle because they know that this method is gonna get a better solution than has ever been seen before. So for example, in all state, the actuararial models, the error of them in predicting who's gonna crash their car over three times better. Oh, wow. Yeah, so on the other hand, there are also companies with no data scientists, not much understanding of modeling who use Kaggle, not just to fill in the gaps, but also to, at the same time, fill in the gaps and also get NASA quality solutions. Interesting. So I noticed your tagline at Kaggle is making data science a sport. Right, isn't that great? I love the tagline, and so my question is, we've talked a little bit, we've covered on theCUBE, and then SiliconANGLE and Wikibon, some of the characteristics needed to be a data scientist. So we know that you've got to be persistent and we want to experiment. I got four. You ready to hear my four? I'm ready. Okay, so the four things that I see in every Kaggle competition winner are as follows. Open-mindedness, creativity and curiosity, tenacity, and deep skills. Now, imagine you didn't have one of those skills. Just one. So maybe you're not tenacious, or maybe you're not the most tenacious guy out there. When you get past after three weeks by someone and you've put your heart and soul in this and suddenly you're not first on the leaderboard anymore, you'll give up. Or maybe you're not that open-minded. So somebody else who's more open-minded will find that crazy idea that you hadn't thought of. Or maybe your skills are slightly weak, in which somebody else will get that last little bit of the model by finding a more performant way to run it. Or maybe you're not quite kind of curious and open-minded enough. Perhaps you'll be in a situation where you make assumptions which just aren't valid. So those are the four skills that we see. Every week we invite one of our competition winners into the Kaggle office to come and sit with us. We've had guys come in from Ecuador and Australia and all over the world. They are all amazing people that have all of these four qualities. And as a result, they're actually terrific people to spend time with because they don't have an ego. They're very curious, they're very passionate, but they're also very opinionated and very skillful. Interesting that the tenacious part, because when I think data science is a sport that implies competitiveness. So is that kind of what you're getting at there? Is that being a competitive, having a competitive nature important for a data scientist? These people are very collaborative and also very competitive. You know, it's not an oxymoron. These are people who are very curious and very passionate. They want to learn what other people are doing. They want to learn how to become better themselves and they don't give up. So yes, they're competitive, but it's in a very positive and collaborative way. And often the best Kaggle data scientists will team up with other past winners and work together on future competitions. That's interesting because as we're doing more research around this we're finding it's a very collaborative discipline, isn't it? Oh, I mean, because it's the discipline that's above all other disciplines. Everyone uses data. And so we are all people that believe in using data to drive decisions rather than using where you went to school or who talks the loudest or who went to the best university. I think data scientists as a whole believe in meritocracy. That's kind of why we're here. And to me Kaggle is the world's largest meritocracy outside the world of sports. Cool. Well, thanks so much for coming on. Thank you. We appreciate it. For our viewers out there, go check out Kaggle. Really cool company doing some really cool stuff in the data science space. So thanks again for coming on. We're gonna take a short break and we'll be back.