 From around the globe, it's theCUBE with digital coverage of AWS re-invent 2020. Special coverage sponsored by AWS Worldwide Public Sector. Welcome to theCUBE virtual. This is our coverage of AWS re-invent 2020, specialized programming for Worldwide Public Sector. I'm Lisa Martin. Got a couple of guests here from Noveta. Please welcome Steven Adelman, Principal Computer Scientist and Kevin Heald, Vice President of Information Exploitation. Gentlemen, welcome to theCUBE. Thank you, thank you for having us. All right guys, so Kevin, we're going to start with you. Give our audience an introduction to Noveta. What do you, what do you guys do? Who are you? How do you play in the public sector government space? Great, yeah, thank you Lisa. So Noveta, Noveta is a technology services company focused on government solutions. So primarily national security solutions. So think customers such as DOD, the intelligence community, FBI, law enforcement and things like that. About 1,300 employees worldwide, primarily in our field, clear resources that really focus on cloud-first solutions for our customers. So solving the tough mission challenges our customers have. So that can be in technology solutions such as data analytics, AIML, IoT, secure workloads, full spectrum cyber, COP, video processing, really anything that's a high-end technology solution is something we do for the government. We have been a privilege, it's a privilege to be a partner with AWS for some time now. In fact, I think the first reinvent we may have been through, Stephen, was six years ago, five years ago. 2012 or 13. Yeah, so we've been around for a while, really kind of enjoying it. And certainly sad that we're remissing an in-person reinvent this year, but looking forward to doing it virtually. So we are actually an advanced year partner with AWS with a machine learning and government competency and really kind of to pump the ML side of that. That was one of our first competencies with AWS and led by a center of excellence that I have in my division that really focuses on machine learning and how we apply it for the mission. And so really we focus on protecting the nation and protecting our equities as a country. And on behalf of the country, we thank you. Stephen, give me a little bit of information from a double-click perspective as computer scientists. What are some of the key challenges that know that it helps its customers solve and how do you do that with AWS? Yeah, thank you. So really as a company that is data-first, so our initial love and still our kind of strongest competency is in applying solutions to large data sets. And as you can imagine, the bigger the data set the more compute you need, the more resources you need and the flexibility from those resources is truly important which led us very early as an especially in the government space and public sector space to be an early adopter of cloud resources because of the fact that rather than standing up a 200 node cluster at many millions of dollars we could spin up AWS resources, process a big data set and then get the answers an analyst or an operator needed and then spin down those resources when that kind of compute wasn't needed. And that is really kind of informed how we do our work as novatans that cloud infrastructure and now pushing into the edge compute space is still kind of keeping those cloud best practices in play to get to access to more data that the two biggest I think revolutions that we've seen with regards to using data to inform business processes and missions has been that cloud resource that allows us to do so much with so less and so much more flexibly. And then the idea of cheap compute making it to the edge and the ability to apply sensors to places where it would have been operationally or cost prohibitive to do that. And then ironically, those are two things that aren't necessarily data analytics or machine learning focused, but man did they make it easier to collect that data and process that data and then get the answers back out. So that really has kind of shaped a lot of the way Noveta has grown as a company and how we serve our customers. So Kevin back over to you, let's one of the things that we've been talking about almost all year is just the acceleration in digital transformation and how much faster organizations, private sector, public sector need to innovate to stay relevant to stay competitive. How are you working with government customers to help them innovate so quickly? You know, we are very fortunate that a set of customers that's focused is actually innovation. It's focused as I read. And you know, we can't do the cool things we do without those customer relationships that really encourage us to try new things out and quite frankly fail quickly when we need to. And so by establishing that relationship, what we've been able to do is to blend agile development, agile acquisition with government requirements process, right? If you know, the typical stereotype of government work is it's this very stove piped hard core acquisition process, right? And so we've been fortunate to instead try quick win kind of projects. And so one of the biggest things we do is partner with our government customers and try to find a difficult challenge to solve over a six to 12 month time, right? So instead of making this long four or five year acquisition cycle, it's like show me, right? How can we solve this problem? And then we partner with the mission partner, show success in six months, show that we can do it with a smaller part of money. And then as we're able to actually make that happen, it expands in something bigger, broader and then we kind of bring it together a coalition of the willing, if you will, in the government and saying, okay, are there other stakeholders that care about this problem, bring them on, bring their problems and bring them together? And we can't do that with some of the passionate people we have, like Steven's a perfect example. When we talk about Picard and the project we're doing here, Steven's passion for this technology partnered with our customers having these challenges and try to enhance what they're doing is a powerful combination. And then the last thing that we're able to do as a company is we actually spend a decent amount of our own dollars on IRED. So R&D that we fund ourselves. And so while finding those problems and spending government dollars to do that, we also have spent our own dollars on machine learning, IoT, sensor, next gen 5G and things like that and how those can partner together, partner together to go back to the government. Yeah, Kevin. Go ahead, Steven. Oh yeah. So I would even say, there's a conventional wisdom that government is slow in plotting and a little bit behind commercial best practices. But there are pockets in growing pockets across the government where they're really kind of jumping ahead of a lot of processes and getting in front of this curve and actually are quite innovative. And because they kind of started off from behind they can jump over a lot of kind of middle ground legacy technologies and they're really innovating. As Kevin said, with the Picard platform, we're partnering with PEO Digital in the Air Force and SAFCDM and Air Force Security Forces as that kind of trifecta of stakeholders who all want to kind of saw a mission problem and wanted to move forward quickly and leave the legacy behind and really take a quantum leap forward. And if anything, they're driving us to innovate more, to introduce more of those kind of modern back practices. And Noveta as a company loves to find those spots in the government sector where we've got those great partners who love what we're doing and it's this great feedback loop where we can solve hard technical problems but then see them deployed to some really important and really cool and impactful missions. And we tend to recruit that kind of nexus of people who want to both solve a really difficult problem but want to see it executed in a really impactful way as well. And then that really creates a great bond for us. And I'm really excited to say that a lot of the government is really taking a move forward in this realm. And I think it's just good for our country and good for the missions that they support. Absolutely, and it's also surprising because as you both said, there is this expectation that government processes are lengthy, laborious, not able to be turned around quickly. But as Kevin, you just said, helping customers, government agencies get impact within six to 12 months versus four to five years. So you talked about Picard, interesting name. Kevin, tell me a little bit more about that technology and what it is that you guys deliver that's unique. Well, honestly, it's probably best to start with Steven. I can give you the high level. This is Steven's vision. I have to give him credit for that. And I will say we have lots of fun acronyms. So it is an acronym, right, Steven? Doesn't it actually stand for something? It stands for platform for integrated C3 and responsive for defense. And I know that the Star Trek theme is a leg up from the last set of programs I had which were my little ponies, so. Oh, wow. That's a definite step in a different direction. I like it. Part of the great thing about working in the government is you get to name things cool things. But to get to your question, so Picard really sprung out of this idea that I had a few years ago that the world, but for our spaces, the Department of Defense and the federal government was going to see a massive influx of the desire to consume sensors from areas of responsibility, from installations, and frankly from battlefields. But they were gonna have to do it in a way that presented some real challenges that you couldn't just kind of throw compute at it or throw traditional IT processes at it. We have legacy sensors that are 40 years old sitting on installations. Old programmological controllers or facilities control systems that were written in cobalt in the 70s, right? And are not even IP based, most of them. And then on the other end of the spectrum, you have seven figure sensors that are throwing out megabits of second of data that are mounted to the back of jeeps, right? That are bouncing through the desert today, but will be bouncing through the jungle tomorrow. And you have to find all of those kind of and combine all of those together and kind of create a cohesive data center for data set set for the mission for what we call a user defined common operating picture for a person to kind of combine all of those different resources and make it work for them. And so we found a great partner with security forces. They realized that they wanted to make a quantum leap forward. They had this idea that the next defender so they're like a military police outfit that the next defender was going to be a data-driven defender. And they were gonna have to win the information war as much as they had to kind of dominate physical space. And they immediately got what we were trying to achieve. And it was just this great synergy. And then we've piled on some other elements and we're really moving that platform forward to kind of take every little bit of information we can get from the areas of responsibility and get it into your modern data lake where they can extract information from all that data. Kevin, as the VP of information exploitation that's a very interesting title. How are you helping government organizations to win the war on information, leverage that information to make big impact fast? Yeah, I mean, I think a lot of it is is that we try to break down the barriers between systems and data so that we can actually enable that data to fuse together to find and get insights into it. As ML and AI have become trendy topics, they're very data hungry operations. And I think what Stephen has done with the card and his team is really we wanna be able to make those sensors seamless from a plug and play perspective that I can plug in a new sensor. It's a standards-based interface that sends that data back so that we can then take it back to the user to find operation picture and make some decisions based off of that data. What's more is that data could even be fused with more than the data that Stephen's collecting off those sensors. It can be commercial data, other government data. And I think as Stephen said earlier, you have to get it back. And as long as you've gotten it back and you ever share it with some of our mission partners then you can do amazing things with it. And Stephen, I know you have some pretty cool ideas on what we're gonna do on the edge, right? How do we do some of this work of the edge where a sensor doesn't allow us to pull all that data back? Yeah, and to follow on to what you were kind of referring to with regards to handling heterogeneous data from different sensors. One of the main things that our government customers and we have seen is that there are a lot of, historically there are a lot of vertical solutions where the sensor, the platform and then the data lake are kind of all part of this proprietary stack. And we quickly realized that that just doesn't work. And so one of the major thrusts of that card platform was to make sure that we had a platform by which we could consume data through adapters from essentially any sensor speaking any protocol with any style data object, whether that was an industry standard or a proprietary protocol, we can quickly ingest it and bring it into our data lake. And then to pile on to what Kevin was talking about with compute, right? So you have like almost like a Maslach's hierarchy of needs when it comes to cyber data or to this IoT data or kind of unified data, you wanna turn it into basic information, alerts alarms, then you wanna do reporting on it or analytics or some higher level workflow function. And then finally, you probably wanna perform some analytics or some trending or sort of anomaly detection on it. And that gets more computationally intensive each step of the way. And so you gotta build a platform that allows you to both take some of that high level compute down to the edge, but also then bring some of that data up into the clouds where you can do that processing and you have to have kind of fungibility between that. And so that card platform allows you to kind of bring GPUs and high processing units down to the edge and make that work. But then also, and then as maybe even a first pass sieve to rule out some of the most, some of the boring data in the video analytics platform we call it blue sky and blue ocean, right? So you're recording lots of video that's not that interesting. How do you filter that out? So you're only sending the information of the interesting video up. So you're not wasting bandwidth on stuff that just doesn't matter. And so it's a lot of kind of tuning these knobs and having a flexible enough platform that you can bring compute down when you need it and you can bring data up to compute on big cloud while you need it and just kind of finding a way to tune that. That really does, I mean, that's a lot of words about how you do that but what that comes to is flexible hardware and being able to apply those dev ops and CI CD platform characteristics to that edge hardware and having a unified platform that allows you to kind of orchestrate your applications in your services all the way up and down your stack from microcontrollers to a big cloud instantiation. You make that sound so easy, Stephen. Kevin, let's wrap it up with you in terms of like making impacts and going forward we know the edge has exploded even more during this very interesting year and that's going to be something that's probably going to stay as a permanent impact or effect. What are some of the things that we can expect in 2021 in terms of how you're able to help government organizations capitalize on that, find things faster, make impact faster? Yeah, I mean, I think the cool thing we're seeing is that there's a lot more commoditization of sensors. There's a lot more sensor information so let's use LIDAR as an example. Things are getting cheaper and so we can all of a sudden do more and more things of the edge than we ever would have expected. When Stephen's team is integrating camera data and fence data from 40 years ago, it's just saying on off, it's not doing anything fancy. But now Stephen, I can't remember what the measure you gave me before was but the cost of LIDAR has dropped so significantly that we can now then deploy that. We can actually roll it out there and not be locked in there for proprietary system. So I see that being very powerful. Also, I can see where you start having sensors interact with each other, right? So one sensor finds one thing and then a good example that we've started to experiment with and I think Stephen, you could touch on it is using like triggering a sensor triggers a drone to actually investigate what's going on and then therefore a hybrid video back and then automatically can investigate instead of having to deploy a defender to actually see what happened at that end point, Stephen. I know there's some more detail you can provide there. Yeah, no, so exactly that Kevin. So the power of the sensor is something old that gives you very uninteresting data like a one or a zero or an on or off can detect something very specific and then do something kind of high speed like task a drone to give you a visual assessment and then run object detection or facial recognition on, you know, do object detection to find a person and do facial recognition on that person to find out if that's a patrol walking through a field or a bad guy trying to invade your space. And so it's really the confluence and the gestalt of all of these sensors in the analytics working together that really creates the power from very simple delivery. I think there's this idea that, you know, 100 bytes of data is not that important but when you put a million sensors giving you 100 bytes of data you can truly find something extremely powerful. And then when you make those interactions sing it's amazing to us the productivity that we can produce and the kind of fidelity of response that we can give to actors in this space whether that's a defender trying to defend a base or a maintenance person trying to proactively replace of the fan or clean the fan on an HVAC system. So, you know, there isn't a fire at a base or interesting enough one of the things that we've been able to achieve is we've taken maintenance data for helicopter engines and we've been able to proactively say, hey, you need to take care of this part of the helicopter engine and it saves money, it saves down times, it keeps the birds in the air and it's a relatively simple algorithm that we were able to achieve and we were able to do that with the maintenance people bring them along in this endeavor and create analytics that they understood and could trust. And so I think that's really the power of this space. Tremendous power. I wish we had more time to dig into it guys. Thank you so much for sharing, not just your insights, what Novot is doing but your passion for what you're doing and how you're making such an impact. Your passion is definitely palpable. Stephen, Kevin, thank you for joining me today. Well, thank you. For my guests, I'm Lisa Martin. You're watching theCUBE virtual.