Ep. 49: Doug Matty of the Army AI Task Force

Well, this is Jonathan Kersting with the Pittsburgh Tech Council on Techvibe Radio, hanging out with Comcast bring you our Summer of 50 Pittsburgh Tech Stories. As we're winding this series down, this is our 49th episode. And I keep saying as we get down the road on this thing, the interviews are just getting cooler and cooler. And I really cannot wait to bring you this segment today because this is something that's going on in Pittsburgh. It just it completely blows my mind and makes me so proud. This is happening in Pittsburgh, and today we are talking to Dr. Doug Matty, he is the director of the army artificial intelligence capabilities. And I tell you what the Army AI Task Force and what they're doing is just amazing. And so Doug, welcome to the show today. So glad to reconnect with you. We had talked to you on tech five, I think about a year or so ago. And we only had a little little snippet to give folks and so I'm glad we can stretch out a little bit today and learn all about some of the great work that you guys are working on today.

Yep, look forward to it. And thanks for the time. For the opportunity.

Absolutely, absolutely. So first and foremost, let's just start with your background because I mean, your background is pretty freakin amazing. I mean, military cyber security stuff, or I'm sorry, artificial intelligence stuff like all around. You've been working on some crazy stuff. Tell us a little about yourself real quick. And let's jump into what the task force is up to these days.

Sure, so a little bit of background. So as you know, I'm a I'm a native I was born in Pittsburgh raised in Irwin, before heading off to the United States Military Academy, West Point is, was fortunate enough to serve in leadership positions in the army. And then at about the 10 year point switched over and was afforded a chance to go to the Naval Postgraduate School for a master's degree in applied math. And then from there, I was able to get into some really neat opportunities for both in terms of how the army runs in the development of the howtos capabilities and then was afforded a chance to go off for a PhD at MIT, wow, engineering systems. And then from there served in a series of really cool jobs where, you know, basically was the overall systems engineer for all of the these warfighting capabilities, then moved on to Cyber Command to help with their capability development, was exposed to how the army does resourcing with a long range resource flow forecast and its program efforts. And then was able to get into the artificial intelligence work that we started here in Pittsburgh back in 2018. And continue to push the envelope here. So that's a little bit different than most officers recently was selected to come on as what they call a Senior Executive Service member. For folks that aren't familiar with that. It's essentially the civilian equivalent to a general officer as a senior executive. Okay, I assume the role as you mentioned, is the director for The Army's artificial intelligence capabilities, which is not only focused on the work that we're doing here in Pittsburgh, but actually pools from across the nation, right across the government labs, academia, industry, etc, to bring all that together to really help support our warfighting and business efforts.

Wow. I mean, you you have some awesome stories to tell him and then what you probably couldn't even tell me. You'd have to kill me afterwards. Yeah. You definitely got an awesome background and how exciting that you're taking all this experience. You have read this up here in Pittsburgh. So what are the what are the nuts and the bolts of what you guys are doing here in Pittsburgh, and you're partnered up with Carnegie Mellon, which is just obviously the powerhouse for all things, artificial intelligence, then we can really start nerding out on all the different ways that AI is being used in the military, which I think is so amazing.

Yeah, so um, so I guess the best way to set the stage is In 2018, the army stood up. Its its newest forestar command, and headquarters called the army army futures command. Okay, and so that was that was established in Austin, Texas. And what it really grew out of was this effort to really push the technical logical boundaries by pulling together all the different aspects of what we call modernization. Okay, so they were focused on things like, how do we generate long range effects? What's the next generation of vehicles that we can use on the battlefield, you know, future vertical lift, you see different renditions of, you know, different prototypes and those types of things of aerial vehicles. And then there's some enabling technologies that we look at as well, whether it's, you know, a short position navigation and timing with our space capabilities, as well as things like synthetic training environment with augmented reality that he Schools, both training and operation. So these are the cross functional teams that were established to really accelerate and integrate development efforts across the army. And so as that headquarters was stood up, and we looked at what were the premier technologies that would help enable us to move forward in our operational doctrine, we recognize that artificial intelligence was really the key enabler. In fact, our Commanding General, General Murray talks about three technologies. AI is first and foremost, followed by autonomous systems in robotics, and as you know, Robo Berg is pretty much the home for a lot of thought leaders in those respective areas.

Yeah, itcuts across all of that means like the backbone of everything in that regard.

So there's, there's obviously lots of opportunities to leverage it. We don't want to call it the panacea, right? It's not gonna solve everything. But the the advances that we are now realizing with the blending of, you know, technological advancements in terms of hardware and memory and storage and compute, as well as the advances on the programming side, that we're no longer just rules base where you can have vast amounts of data, you know, help train the computer to be able to do the things that we needed to do is just been revolutionary.

Absolutely. And the other thing is, it's changing so quickly as well too. And the advances just keep coming and coming and coming. So just as you get something kind of plugged in and ready to go, you see another bit of changes that now have to be accommodated for. How's that like to be? Yeah, so um, so what my boss tells me is that while everybody else is using a calendar, I need you using a stopwatch. So I think that kind of captures a lot of dynamic that you're describing where we operate in partnership with academia and industry, not on a year long type of effort. But actually, you know, look into rotate through, I'll say, 60 to 90 day sprints, right? We can drive the teams to develop, test it out, put it in the hands of soldiers. Let them wrap it up some give us some feedback, and then do it again. Right.

So you just keep keep iterating as you keep going as well, because the environment just keeps changing, and the threats keep changing, and everything keeps changing. And you're, you've got that stopwatch always going off so you don't bring a dress to you.

And so like you said, you know, not only is it a change in requirements, but also a change in use cases. So, you know, the advances we're making a technology are such that we're finding areas of application that we hadn't even realized were possible. I mean, a great example of that here in Pittsburgh is if you think about it, you know, folks over at Pitt that work with the medical field have been using evidence based Based Medicine and empirical methods for diagnostics and healthcare, etc. in you know, for those folks, they it's a natural segue that they go from medicine and the body as a system to predictive maintenance on our, you know, our platforms and systems that we operate in the arm. And so there's a good synergy. Back to your point about where we can learn from things that don't necessarily seem to be related, but actually, you know, really cause us to advance quickly.

Absolutely, that I just find it just just so exciting with all that work. So I'm assuming so when you set this thing up in Austin, you're you're then looking for a center of excellence within artificial intelligence. You see Carnegie Mellon, you see Pittsburgh, and you're like, this is where we got to set up so we can be close to this to this resource. Is that how that worked out?

Yeah. And so actually, it was part of a conversation at the VOD level department defense late level at The same time that we were looking to stand up, our our AI organization, the d. o. d was standing up what's known as the joint AI center. And so when we had the discussion with the army leadership and God leadership, they said, Look, we need to find somewhere that has not only the breadth, the depth, to be able to engage with us. And so that's, that's partially why, you know, looking at Pittsburgh, we were able to find that because, again, you have a thought leadership here in academia, you have the innovative, hard working spirit of, you know, both this, this ecosystem that is in Pittsburgh with robotics row, a number of members that you have at the pet Council was, you know, there's just a really vibrant dynamic that's going on as well as some of these companies that were startups have now been acquired by larger right, tech based companies that now have a footprint here in Pittsburgh just because of the goodness that that you can grow Hear very much So tell me about something Tell me about your your footprint here. So you guys are what you're based out of vn rack? Is that what it is?

Yeah, so um, so initially what the taskforce we were split between Pittsburgh and in the National Capital Region. Okay. And then based on the progress that we had here you know, the army leadership saw the saw the benefits of having everybody co located in Pittsburgh. So about a year after we were stood up the direction came to move everyone from DC up to Pittsburgh. Once we got the team together, as you can imagine, things started really clicking, then all those cylinders are firing, etc. And so, March right before all of the remote working started due to, you know, the current health situations. We had a visit again from the army leadership that looked at what we were doing in the direction we were going and said, Look, I I believe in what you're doing. You're doing great work on one, I want to expand the scope. And so we really went from just, I'll say, a basketball team worth of around 1520, folks, etc. Basically tripled our size. 

Yeah, so, so we're really gonna be excited about getting all those folks here. And what's unique about the organization is, when we were first it up, we were very much just focused on the project work, and what we in partnership with our teammates here in Pittsburgh to develop. And so now that we've achieved success across the board on those four thrusts, and I'll talk about those in a second. We laid in the other aspects of, you know, this AI group that allows us to work, you know, requirements and capability development across all of the army enterprise, right. So we have folks that are, you know, typically you don't find this In one organization where you have folks that are doing development, and the folks that are generating the requirements, and the folks that are working with the industry partner, to pull in all those capabilities, whether it's again, industry or academia. So with having that all as part of one group, you know, we've realized the synergy, and now we look to expand and scale that, that effort, right? I mean, what's that, like tripling your staff, like in like, like, a year? I mean, that's just really cool.

Yep. So, so the good news is that, you know, we have the direction to plan that out. We started that in March, you know, was able, we were able to get approval for that here late in the summer. And so now we're actually beginning the, the fun part of it, which you just alluded to. And so again, we're looking, you know, there's a lot of talent that's here locally. And so we'll be putting out information and hopefully find some of those folks that are here that want to be on the team as well. As I draw from across the nation, really there's a lot of folks with a recent event that we had when I was sworn in to take over as director. I've had engagements from folks all across the country asking about how do they How do they get on the team? And I tell them, you're on the team. So you say you're not so come on and get on the team. So we're looking at Pittsburgh, very clear, can bring a lot of super smart women and men to Pittsburgh working for you guys. That's, that is very exciting. to can you kind of maybe kind of go he's kind of you'd be like, he's like four areas you were mentioning about how you guys are operating?

Sure. Yeah. So um, you know, I'm not sure if everybody realizes, you know, exactly how broad an enterprise the army is. And so of course, they, they think of our main focus, which is our operational capabilities as we secure and defend our nation, but it takes a lot behind that to support that. So you know, what I tell folks when I talk through the projects is First and foremost, most important in the army are our people. And so there's no surprise that when we were set up, the Secretary of the Army said, I wanted to use all of this advanced capability to help us with our talent management of our most valuable resource, which is Rp. Right? So we've been working through that the develop our different algorithms and systems, etc. to kind of identify those hidden kind of implicit characteristics so that we can maximize, you know, folks opportunities, our soldiers, opportunities and civilians, you know, to contribute to the team. Right. The second, the secondary that we did, as you can imagine, as I mentioned, DMC was standing up their effort and so as they were getting their feet under them, you know, a lot of times they'll work with a services to be what they call executive agents, because we have, you know, the staff, we have the processes, we have the institutional support to carry on with these types of efforts. And so, what we did was we stood up a group focused on pretty addictive maintenance, which is really kind of our inroad into the broader logistics and support structures that that we have. So as you know, when you think about our, the our military, the key thing that differentiates us is our ability to project power. And that really comes from our ability to sustain in which the army is recognized as the executive agent given the number of systems we have soldiers etc. So, it was a natural fit for us to to get into that effort and start off with predictive maintenance on how to do that is part of it as I mentioned before with the establishment of Army futures command, we the army took a hard look and said, Look, we need to prioritize and kind of go after we can't do everything at once, so we need to prioritize and so the next two areas that we are focused on are right in line with the army modernization effort. So as I discussed earlier about our ability to generate long range effects, but precision. One of the hardest challenges to that is how do you understand and have battlespace awareness in those types of distances. So what we did was we had folks that had been partnered with what we call training with industry opportunities, a broadening assignment for our officers. And so they are embedded in industry partners and work right alongside with the other employees of the company to learn, you know, and be exposed to different application skill sets that we may not have already in the army. But we know his potential that there may be a need for that in the future. And so it's a it's somewhat akin to a fellowship, if you will, where you send someone off for a year and let them just be exposed in an environment which they may not be. And so, you know, we were fortunate that one of our officers had been working with the commercial partner that did a lot of work with what they call it orbit satellites. And so as they go along and take pictures of, you know, farms and agriculture and soil in different, you know, weather patterns or those types of things, you can imagine this picking up a number of other things that would be interesting.

Absolutely. Yeah.

Yeah. And so that was a natural fit, as you mentioned before, about where can we apply artificial intelligence? Well, you know, rather than having, you know, someone that an analyst sit there and look at frame after frame after minute, after minute, etc, we're able to train the model to say, look, can we identify some of these key things that we're looking at whether it's, you know, road construction, whether it's just changing infrastructure, you know, specific types of equipment and such. You know, we're able to leverage that and so what we call that as our Intel support for operations for long range precision effects, which is again, our number one priority number two, is really where you're you really glad to be in Pittsburgh with the second one because Next Generation vehicles is really a natural extension of the autonomous vehicle efforts that are in and around town here because based on their ability for perception, to be able to navigate the complexities of an urban environment on the road, you can imagine how it would be whether we're in an urban environment, or a rural environment or desert environment, etc. and the ability to navigate and have what we call operationally relevant maneuver is critical. And then, you know, the other part of that is not only the maneuver aspect, but can you have protective sensory around you so that you understand who's around you, what's around you, what are they doing, what are the actions, who's on your left, who's on your right, those types of things. So, what we call there's automated threat recognition, which is really an enhancement to protect our soldiers and enhance their mission capabilities. So those are the four thrusts and now the good news is I've mentioned before We started really in earnest with those efforts, I'd say, spring of 2019 in, we have actually delivered capability across all four thrusts. They've been implemented into other systems where we were able to integrate those in or in fact, in some cases, generate new systems that are now supporting our army's efforts. So the fact again, we were able to do that in less than a calendar year is really what caught notice of the leadership and why I think they're so excited about, you know, the progress that we're making here.

They have to be I mean, that is just astounding to get all to make you progress the way you have across very complicated things. And this is this is not easy things you guys are attacking here, really goes to show you just the effectiveness so far. And we also talk about like I thought this was really cool, got this press release, like a month or so ago about the inaugural class of artificial intelligence scholars that you guys are working with as well too. I mean, how cool is that?

Yep. And so again, this kind of goes back to our model. You know, it's, it's customary for folks to think of defense. modernization is where folks inside the VOD will come up with the good ideas. And then they outsource it to, for all the right reasons, the industry to go and develop and build the challenges that, you know, there's, you lack that virtuous cycle of learning about the technology, which then as we talked a little bit earlier about, expands your your trade space of what those use cases could be used for. Right. And so what we wanted to really do was not only build the systems but also build the batch. And so it was with, you know, our efforts to develop the workforce. So if you look at the results from the recent National Security Commission on artificial intelligence, which was sponsored, you know, and the auspices of I think it was charted by Congress, but executed, obviously, under the executive branch, what they came back and talked about in the most, you know, vibrant area where there's an opportunity to improve the most was with the workforce. And so the one thing the army knows how to do is we know how to train folks, right? We know how to, yeah, that's exactly what we want to do is we kind of wove in this workforce development effort along with the projects. But then we also wanted to augment that and have several lines of effort to accelerate that learning in terms of our workforce. So I mean, you know, is we think about how best to do that with the army that obviously the key thing is you have to develop leaders, right. That's, that's, that's really what the army is focused on.

Right, comes back to the people.

That's right. And so, so we we went to our academic partners, hold on The shelf what we saw was, you know, kind of the premier training and education effort that they have for leaders. And so we started that pilot up, but first focused on enterprise data management. Because you can expect when you're trying to establish AI at scale across the enterprise, it really is essential that you have a solid foundation in terms of the data. And so we started with that and completed that first course with a number of general officers and senior executive, senior executives. And then the second course is now we're rolling out with the commands where we're pulling in teams from the respective commands, whether it's Army Materiel Command that's in Alabama, whether it's our Training and Doctrine Command in Fort Eustis, Virginia, I mean, it covers all of those major commands. And what we'd like to do is have them go through and learn as leaders, that next level of detail so that it's not just enough to spell AI but they actually under And what are the leavers that have to be pulled? What are the policies that have to be in place? What are the dynamics that they, as the leaders have to ensure are, are being enacted so that they can successfully leverage data driven leadership? Because that's really what AI does is enabling our soldiers and our leaders on in their mission set. And so that's why we're focused on data driven leadership. Yeah, so again, that's the leadership aspect. But then, again, while they are setting the stage, and, you know, supporting the efforts, you have to have that professional workforce that can do it. And so that's where we started off with our, what we call AI Scholars Program. And, again, it's it's really difficult to find an individual that can do all of the various aspects needed to do AI across the enterprise. And what we recognize is, it's okay to have folks kind of they're still overlapping But just as we would have, you know, teams that man our systems, the same thing comes to our data systems and our AI capability. And so what we've identified are, we're gonna have a number of folks that are focused predominantly on the data analyst portion, which is the folks that generate the algorithms, right. And then folks that are focused on the data engineering side to ensure that the data is flowing across the pipelines, if you will, to feed into those algorithms and support the whole what we call the AI stack of capabilities, all the way from sensors up through data management to algorithms development, the human machine interface and or economy. And so that's the scholars we paired it up. We start off with, you know, a number of folks that went to the data analyst side and the revenue quarter corresponding number that went into data engineers. They'll complete a master's degree. And then in 22 months after they finished these world class programs, like Carnegie Mellon then they'll be put into assignments for the utilization tour and help develop these efforts. And again, you know, the key thing here is, you know, the number of books that are in the initial pilot is really, you know, just kind of that first start, because we're going to grow the number of folks that are participating, and it has both military and civilians. So again, if you know, one of the key things that we identifies there's, there's a strong component to this war civilian employees are able to, you know, they have the skillsets they may not have been, what they went to school for what they developed them just because that was needed in their current position or jobs. And I think that's common with what you see in industry as well. Definitely folks say, you know, I started off as an engineer, but now I'm, you know, a data scientist or data engineer.

Absolutely.

So we want to afford them the opportunity. So they're in that that cohort as well. And then the idea is just we scale that program here. Carnegie Mellon, watch this scale across other partner schools. So whether it's Pitt, University of Southern California, University of Texas, Austin, I mean, all of those other schools are excited to be part of a team.

They gotta be totally stoked, because you think about I mean, it's like, I think a lot of people don't understand that, like, you can get these skills and that you can actually apply them in the army like you, this is something you can do to like, help your country and have me really work on the leading edge of technology, while also working for the army as well, too, which I think is just this cool combination of being able to kind of like serve your country and have RC cars and work on the moon's leading edge technology going I mean, what's cooler than that?

And the good news is, I think that the, you know, folks in charge of our government in VOD recognize as well. So, what they've done in recent, you know, passages of laws as they've given special authorities that allow us to in No unique situations and circumstances, reach out and touch those folks. So just as the cyber community initiated, where they could have what they call direct commission, meaning they could reach out into the civilian sector and find some of these just amazing talent. Or there's also other ways that we could go about. We can bring civilians with direct hire authority and just expedite getting them on board as well. And some pretty neat opportunities for them to continue to work as well. So, again, it's a choice that we can afford folks that have these special talents and skills that they can be on the team, whether it's in uniform or as a department, army civilian. You know, last spring, we had hoped to have a major event here in Pittsburgh to kind of engage with the broader ecosystem. Of course, we had to postpone that because of the situation but Looking forward to this April, I think that we'll find that there'll be an opportunity for us to have not only the chance to kind of show some of the progress we've made on the project size, but also have these kinds of engagements with the talent that are talent that's around here in Pittsburgh, and bring them on the team in one way or the other.

That's why you got to work with the Pittsburgh tech Council. And that happens, we'd love to be a part of that and help get that out to the community because we think this is such transformative stuff. And everybody needs no, this is kind of going on, because this is something that is really making Pittsburgh great. You guys have your bases covered on this. Anything that we haven't covered, because there's so much going on there. Well, anything else that we haven't talked about that you think is really important for our listeners here to know about?

I think, as I said, we're really just getting started. In the first year that we were here, as I mentioned, those cross functional teams. They had come in, they kind of laid out their problem sets for us to kind of digest on it and find ways that we could adjust Going forward, I think what you're going to find is, in addition to those eight cross functional teams, you'll have a much broader set of cross cross section across the army of folks that that have been working in a number of years to resolve some of these capability gaps. And in addition to that, our sister, our sister services and other elements of God will be here as well. And so I think, to your point, it is key that, you know, these organizations that have kind of helped form these communities of experts in these areas of technology, I think they'll play a key role. So I really look forward to the path forward where we work with the council. There's a association of the United States Army chapter here in Pittsburgh. They kind of act as a conduit to tell the army story but then also get those place where they can go to and ask questions and be involved. And then of course, we We're here with our partners, like I said, with Carnegie Mellon, University of Pittsburgh, and then a number of, you know, industries that already been doing a lot of work with the DMV. And so and I know several have already approached about where they can find space because they want to be part of the action here, Pittsburgh, so it'll be really neat to see where this grows over the next year or so. And I expect it to be pretty quick in terms of how it matures expands.

I think your velocity is tremendous right now it's only to get bigger and better and to me it's just so impressive to see this happening. And happening here in Pittsburgh. So I'm so pumped. You got to tell us a little bit of the story what's going on here and I can't thank you enough. Dr. Maddy to be part of this. It's my everyone. We're talking to Dr. Doug, Matty, the director of the army AI capabilities. And what a fascinating story. We had to keep you coming back on a regular basis. I want updates, because I know you guys are moving and grooving and there's always something new fun to talk to. So glad you got to hang out with me and Comcast be part of our summer of 50 Pittsburgh tech stories you are the best.

Transcribed by https://otter.ai