 Live from Las Vegas, it's theCUBE, covering AWS re-invent 2017, presented by AWS, Intel, and our ecosystem of partners. Hey, welcome back everyone, live here, this is theCUBE in Las Vegas for AWS, Amazon Web Services re-invent 2017, our fifth year covering the event, wall-to-wall coverage, three days out of day two, 45,000 people here, developers, and business connecting together this year, big show, Amazon Continuous Growth, I'm John Furth, my co-host, Justin Moore, and our next guest is from Washington County Sheriff's Office, using Amazon, Amazon recognition, Chris Adzma, who is the Senior Information Systems Analyst at the Washington County Sheriff, welcome to theCUBE. Nice to have you. So Chris, be here. So, tons of cool stuff we saw on stage today. You know, they've had Polly Lex out for a while, we're going to start to see some of these multimedia services around human identification, transcription, recognition's been out for a while. With the power of the cloud, you can start rolling out some pretty cool services, you have one of them, talk about your solution and what you guys are doing with it. Sure, about last year when recognition was announced, I wanted to provide our deputies at the Sheriff's Office with a way to identify people based on videos that we get from either surveillance or eyewitnesses. So I looked into recognition and decided that we should give it a try by giving all of our booking photos or mug shots up to the cloud for it to be indexed. So that's what I did, I indexed all about 300,000 booking photos we have in the last 10 years and put that into a recognition collection and now I can use the simple tools that AWS gives me to search against that index for any new image that we get in, either from surveillance or an eyewitness, allowing us to get identification within seconds as opposed to having to go through all 700 employees at the Sheriff's Office for the chance that they might have known. So the old way was essentially grab the footage and then do the old mug shot, kind of scan manually, right? Yeah, manually. It wasn't in a book, it was on a website, but essentially, yeah, you had to. I made my point, it sucks, it's hard as hell. Very difficult, very difficult. You see on TV all the magic, all the pictures going on and the facial recognition, you see on the movies and stuff. Closer we to that right now in terms of that capability. As far as facial recognition goes, it all depends on the data that you have at your fingertips. Right now I have booking photos, so I can identify people with a very high level of certainty if they've been in our jail. If they haven't been in our jail, I obviously don't have much of a chance of identifying them, so what you see on the TV where it's like, we look through all the DMV records, we look through all of the people on the street and all of this stuff, we're pretty far off from that because nobody has a catalog of all those images. The corpus of all the pictures, all the data. Yeah, but when you have the data, it's very simple. Right, and it's a lot like scanning for fingerprints as people would have seen that, you have a fingerprint that you've collected from a crime scene and we see it on your NTIS or something, scan through all of that. So it's pretty similar to that. It's similar to that or DNA or anything like that. If you have the data set, it's very easy to search for those people. So faces are no different. So how long did it take you to get it up and running? Did you have to ingest the photos? How did you do that? So they're on websites, so you had to add them on digital already. Yeah, from never knowing anything about Amazon Web Services to a fully functional prototype of this product took me 30 days. Wow. I had the photos uploaded and the ability to actually run the searches via the API in three. So extremely easy, extremely easy. So given the success that you've had with that particular product, are there other services at AWS that you're looking into that say, hey, that would actually be really useful for us? Yes, a couple of other announcements today. First off, the recognition for video. Something that we have a problem with and I'm hoping recognition for video is going to help with is when you have a surveillance camera, people are moving all the time, therefore trying to get a screenshot is going to get a blurry image and we're not getting good results with low light or low frame rate, but recognition for video is going to be able to take that movement and still look at the face. Hopefully we're going to be able to get a better facial identification that way. Another thing that I want to look into is this deep lens they just announced today. Awesome. Looks extremely promising in the way of me being able to teach it things that we need. A great example of what I would use this for is when a inmate comes in, we take pictures of scars, marks and tattoos. That way we have a database of all the scars, marks and tattoos on somebody in case if they recommit a crime and an eyewitness says they had a skull tattoo on their chest, we can then look through all the people who have a skull tattoo and say, these are our list of possible suspects. The problem with that is, is that you may enter somebody in as a skull and you may enter it in as crossbones. Somebody else might put an accidental eye in there. So it's very hard to do a text search against that. But if recognition were to come through, or it wouldn't be recognition in this case, if whatever model I built with the deep lens came through and said this is a skull and this is the word we use, then I'd be able to index all of those images, quickly pull them up, so we wouldn't even need a picture, we would just need to know from an eyewitness that there was a skull on that person's chest. We had a guest on yesterday from Thorne, which Intel's doing AI for good, and they used essentially, they didn't say Craigslist, but it's trying to look for women who are being sold, prostitution and exploited children and whatnot. And it's all machine learning and some natural language processing. When you look at the SAGE announcement, that looks promising, because Amazon's trying to democratize the heavy lifting around all this voodoo machine learning. Which I mean, if you're totally a computer science geek and that's all you do, yeah, you could probably master machine learning, but if you're a practitioner, you just whip it up. Well, yeah, and that's a good example because I am not a data scientist and I have no idea how this stuff works in the backend, but being able to utilize, stand on the shoulders of these giants, so to speak, is allowing people like me who, hey, I only have seven people on my team to devote to this kind of thing, we don't have a lot of resources, so we wouldn't be able to get a data scientist, but opening this stuff up to us allows us to build these things, like this facial recognition and other things based on machine learning, and ultimately keep our citizens safe through the work that AWS does in getting this to us. Yeah, we've been saying a couple of different interviews so far that humans don't scale, so these tools provide the humans that you do have a lot more leverage to get things done. We were talking just before we started recording that these are tools that assist the humans. You're not replacing the humans with machines that just go, oh, we're going to seed all decision-making to you. This is just another tool, like being able to fingerprint people and search for that. It's one more way of doing what the standard policing that you already do. Exactly, and the tool that I've already created and any tool I create after that doesn't ever look to replace our deputies or our detectives. We give them things so that they don't have to do the things like flipping through that book for hours upon hours. They can be out in the field following the leads, keeping the community safe, and apprehending these criminals. They have body cameras too? Not yet. We are currently looking into body cameras. That's a trend, they're going to be instrumented basically like warriors, fully loaded cameras. Yeah, I tend not to think of it like that only because, again, that's a tool that we use not to be that land warrior, so to speak, but more of a... Documentation, I mean, you see them on cars when people get pulled over. Exactly. You got the evidence, the video. It's documentation just like anything else and it's just that one more tool that helps that deputy, that detective, that police officer get a better idea of the entire situation. Maybe I shouldn't have said war. Maybe I'm into the Twitch culture where they're all geared up with all the gear. Okay, so the next question for you is, what's your vibe on this show? Obviously, you have great experience. We're going to Amazon, you're a success study because you're trying to get a job done, you got some tools and making it happen. What's your take this year? What's your vibe of the show? I'm really excited about a lot of stuff I'm seeing at the show. A lot of the announcements seemed like they were almost geared towards me and I know they weren't obviously but it really felt like announcement after announcement were these things that I'm wanting to go home and immediately start to play with. Anywhere from the stuff that was in the machine learning to the new elastic containers that they're announcing to the new Lambda functions that they're talking about. I mean, just all over the board, I'm very excited for all these new things that I get to go home and play with. Yeah. What do you think, Justin? What's your take on the vibe show? It's an interesting show. I'm finding it a little different than what I was expecting. This is my first time here at AWS re-invent. I go to a lot of other trade shows and I was expecting more like a developer show, like I'm going to KubeCon next week and that's full of people with spiky hair and pink shoes and craziness. Here it's- That's at the Aria, by the way. Yeah, that's at the Aria, right. It's a bit more casual than some of the other more businessy sort of conferences. I mean, here I am wearing a jacket. So I don't feel completely out of place here but it does feel like it's that blending of like business and use cases and the things that you actually get done with it as well as there being people here who have the tools that they want to go and build amazing new things with. Right, right, yeah. Just that nice blend, I think. Yeah, I found that it definitely doesn't feel like any other developer conference I've been to but being in the public sector tend to go to the more, you know, business suit conferences. So- This is like total developer for you, public, from a public sector perspective. From where I'm coming from, this is very laid back and extremely- Oh yeah. But at the same time, it's very like a mixture, like you said, you see executives mingling with the developers and talking about- You're a good example, I think, of Amazon. Just, I mean, first of all, there's a, the builder thing in the area is supposed to be pretty cool. I was told to go there last night, people came back, it was very much builder, kind of maker culture, they're doing prototypes, it's very, very developer oriented. But the public sector, I'm astonished by Amazon's success there because the stuff is easy and low cost to get in. And public sector is not known for its agility. No. I mean, it's just music to your ears, right? I mean, if you're in the public sector, you're like, what, now I can get it done? Very much so. And one thing I love to share about our solution is the price, right? Because I spent $6 a month for my AWS bill. That's extremely easy to sell to taxpayers, right? It's extremely easy to sell to the higher ups in government to say, I'm going to tinker around with this, but even if we solve one crime, we've already seen a return on our investment above and beyond what we expected. Yeah. No brainer, no brainer. Chris, thanks so much for sharing your story. Really appreciate you congratulating us on your success. Thank you. And keep in touch with theCUBE. Welcome to theCUBE Alumni Club. All right. Very good for coming on, this is theCUBE here. Amazon reinvent, bringing all the action down, all the success stories, all the analysis is here. I'm John Furrier with theCUBE. More live coverage after this short break.