Mission Modernized: How AI is Reshaping Intelligence from the Inside Out | The Pair Program Ep86

Speaker A: Welcome to the PEAR program from Hatchpad, the podcast that gives you a front row seat to candid conversations with tech leaders from the startup world. I'm your host, Tim Winkler, the creator of Hatchpad.

Speaker B: And I'm your other host, Mike Gruen.

Speaker A: Join us each episode as we bring together two guests to dissect topics at the intersection of technology, startups, and career growth. Welcome back to the PEAR program. I'm your host, Tim Winkler. Joined, uh, as my, uh, co host, Sean Leahy. Sean, how's it going?

Speaker B: Well, Tim, it's getting cold outside, uh, which means I'm spending more time inside. Um, and so I'm just going to go straight into my pair me up and jump the gun a little bit here.

Speaker A: Okay.

Speaker B: Um, when winter comes around, or we're getting into fall now, um, like I said, I'm spending more time inside and so I always had the tendency to redecorate. So I've been throwing up some new art, as you can tell behind me. So my pair is hanging art and buying spackle from Home Depot, because you're going to get. You're going to get some, some holes from what, your hangers that were errant and where you measured once and have to cut twice or whatever. And, uh, I'm definitely going to spend some time this weekend, uh, repairing some holes behind this beautiful arc you see behind me. So there you go. When you're hanging art, you're going to be buying some spackle from. From Home Depot.

Speaker A: Absolutely.

Speaker B: Yeah.

Speaker A: And you're doing it indoors, so it's perfect timing for it all.

Speaker B: So that's what I say. Yeah.

Speaker A: And again, Kuda is on the alignment. I think we, we had this conversation, uh, uh, on a previous episode, but those are flawlessly, uh, hung.

Speaker B: Laser levels.

Speaker A: Laser levels, that's right. Um, cool. Well, I'm not going to give you my pairing just yet. I'm going to save it. Um, but I appreciate you. I appreciate you.

Speaker B: I went. I went out of order on that one. So. Yeah, that's on me.

Speaker A: So I, I've got a random question for you. Um, I don't know if you. You saw this. It's a little, uh, a little bit quirky, a little sci fi. Um, Tom Brady cloned his family dog. Did you, did you hear about this?

Speaker B: Well, I did not hear about this.

Speaker A: Yeah, he cloned Colossal. Yeah, Colossal, um, cloned his family dog that he had had for years and then wanted to make sure that his family would have, uh, the ability to kind of experience that dog. And so he cloned it, and, and uh, I guess this was like 2024. So my question for you is, you know, if you could clone one, uh, pet from your past, would you, would you do it?

Speaker B: All right, so the, the obvious leader there is my, my childhood dog, Windermere. That's right. Classic traditional dog name. Um, would I, would I clone him and bring him back or not bring him back? I guess we're talking about a duplicate.

Speaker A: That's right.

Speaker B: Uh, I don't know. That's something I'm gonna have to noodle on. I can tell you an animal that I would not clone, which is, uh, I have, uh, an ex girlfriend I was living with and she had a cat. And that cat and I got along very poorly. So, you know, no ill wishes on that cat, just in case, you know, the cat's listening, but probably, uh, wouldn't want two of those running around.

Speaker A: Yeah, it's interesting. Uh, I'm not really on the, uh, cloning train. Um, but it kind of freaks me out. But I will say that knowing how obsessed people are about their pets, I can see people kind of getting behind this, but it's a little, uh, weird.

Speaker B: Well, there's the Y Combinator 2027 class clone your pet. That's probably going to get a seed round for sure.

Speaker A: Yeah, we're going to have to get them on. Uh, today's episode has nothing to do with cloning at all. Uh, so this, that's an awkward segue into today's episode. Uh, today's episode is about modernization in the intelligence community. Uh, and specifically how artificial intelligence is changing. How intel is collected, analyzed and used in real time. Um, to break this down, we've got two excellent guests. Uh, Bruce Frost. Ah, VP of Intelligence at Rhombus Power, a startup that delivers real time AI powered decision intelligence for the world's most complex missions. Uh, Bruce is a 30 year veteran of the intelligence community now focused on advancing AI to modernize how the mission gets done. Bruce, thanks for joining us.

Speaker C: Happy to be here.

Speaker A: And then alongside Bruce, we've got Sheetal Patel. Uh, Sheetal leads the CIA's or led the CIA's push to harness emerging technologies for national security. Serving as the Assistant Director for the Transnational Technology Mission center. Brings, uh, that expertise to the private sector now and guides companies at the intersection of innovation, intelligence and risk. Uh, both of you all, thank you for, for joining us on the PAIR program.

Speaker D: Thank you for having us.

Speaker A: All right, now before we do get into the heavy stuff, we kick things off with a fun segment we call Pair me up. We all go around the room, we spitball two things that just go good together. Sean, I'm not going to start with you because you started with yours. Uh, so I'm going to jump in and, uh, kick things off with. With my pairing. So, um, from the D.C. metro area, uh, in Virginia, we've been dealing with some extreme, uh, weather, um, and had a pretty much Arctic cold blast hit us. Um, and so I'm going to go with extreme weather and H Vac mishaps. Um, so at the time of this recording, um, yes, we had temperatures go from, like, it felt like 70 degrees out this weekend, uh, down to basically the. The low 20s, mid, uh, 20s last night. And so, of course, the timing of this is ideal. My H Vac kind of woke up. It was 56 degrees in. In my house, and I'm sitting here trying to phone call different contractors to come out, help troubleshoot this. And, um, we got it fixed, uh, fairly quickly, but this isn't the first time, like, this is. This has happened. Uh, it.

Speaker C: It.

Speaker A: It's happened in, like, the past couple of years where you have, like, the extreme heat spikes and then the AC is just kind of conking out. So I think I'm just, uh, permanently cursed when it comes to extreme weather spikes. And then, uh, H vac mishaps. So that's my pairing. Um, just fortunate we got it fixed because we woke up and we were just chilled this m. Morning.

Speaker B: Tim, I'm. I'm sensing a theme here, right between your H Vac mishaps and some of my, you know, interior decorating mishaps. And I know we've been. We've been collecting hypothetical, um, sponsors on the past couple episodes of the Pear program, so. Home Depot, Lowe's, you know, if you guys want to come in here with a sponsorship, I think we're all about it on the pair. You know, just. Just hit us up. But, uh, we could definitely probably work something out there.

Speaker A: Yeah, big, Big home renovation guys here. So whatever. Whatever you need. We'll get your project done. Uh, awesome. Let's pass it over to our guest. Uh, Bruce, uh, what kind of pairing did you bring to the conversation?

Speaker C: So I'll tell you. My favorite pairing, at least right now, is coffee. Because my son runs a coffee business and cycling. And with the cold weather here, I moved all my cycling indoors. Got my. My psych. Set up my bikes up on the trainer and.

Speaker A: Very nice.

Speaker C: Yeah, can.

Speaker A: Can we get a plug on the. On the coffee shop apiary coffee?

Speaker C: They do high end uh, coffee roasting. And they're online.

Speaker B: I. I've. I've run into them before. How do I know that name?

Speaker C: I don't know.

Speaker A: What's it. What's it called again? Can it. Can. Can you say it again?

Speaker C: Apiary. Like, with the bees.

Speaker A: Uh, okay. Very cool. So I hope you get, like, a family kind of discount, too, when you go there. That's great. Um, awesome. Yeah. Well, the coffee. Yeah, a hot coffee and a bike right now sounds like it's almost necessary.

Speaker C: Well, the travel I do is the only way I survive.

Speaker A: Well, I'm excited to get into some of that on, uh, the main discussion. And then, uh, she told us, uh, your intro, uh, on your pairing.

Speaker D: Oh. So this will not be a surprise to Bruce, but it would be wine and cheese.

Speaker A: Very good. What. What kind of wine?

Speaker D: Red Cabernet Sauvignon.

Speaker A: Okay, Nice Cab. And, uh, any specific cheese you want to shout out?

Speaker D: Ooh, let's see. Gouda, uh, cheddar, Anything but blue cheese.

Speaker A: Do you ever get out to, uh, some of the wineries out towards, like, Western Loudoun? You've been out to Bluemont?

Speaker D: Yeah, I've been to Bluemont. I've been to a couple of others out there. And then there's an Echelon wine bar where a friend of mine teaches.

Speaker A: Yeah.

Speaker D: Phenomenal.

Speaker A: That's great. Yeah, I feel like it's, uh, it's almost a requirement if you're going to go to, like, a winery or something. You got to get some level of, like, a charcuterie, some sort of a cheese with that wine.

Speaker D: Absolutely.

Speaker A: Good stuff. Yeah. Great pairings all around. Uh, let's, uh, let's jump into the. The heart of the discussion. So I'm gonna kick things off and kind of start at the core. Uh, Bruce, from. From your vantage point at Rhombus, you know, what's the kind of fundame problem that you're solving for in the intelligence community today?

Speaker C: Um, you know, I've always heard, my entire career, I heard that open source was the fundamental source of intelligence. We should always be looking to first. And there's a lot of reasons behind that. One is it's ubiquitous. Um, but the challenge always has been, is there's. You have hundreds of languages, foreign languages, you have petabytes of data. How do you actually access that in a way that's usable, uh, and useful for intelligence analysts, uh, in their work? And I think what we've seen the last decade or so is the ability to access that data and then using machine learning models. To make sense of it. So what we're doing is providing, um, really it's a global intelligence platform to our partners and clients around the world so they can see what's happening around the globe and then use that to direct their own collection as needed. But it's a baseline assessment that we provide, um, that I find is rock solid. Um, it's really good analysis that we're producing, uh, based on, um, this global look that we have, all based on the open source. So fundamentally for many of our partners, we're giving them a global look that they don't have otherwise.

Speaker A: Wow. I want to talk deeper about, uh, the product and a little bit on how that global look comes together. But, um, I'll pass it over to you, Sheetal. What was your perspective kind of with inside the ic, uh, what kind of gap stood out to you the most between like the tools that you had and you felt like what was needed?

Speaker D: So I think what Bruce is talking about on the open source side, right, and just the explosion of data and being able to make sense of the data as, as a country. And we're not the only ones around the world, but I think we collect so much data and sometimes we don't know what's in the data and really getting a handle on what we already have and what we know was absolutely critical. And Bruce and I had talked about this when he was inside. We talked about it when he left. I talked about it with others. I think everybody understands this is a huge problem and this is where AI can really help with the large language models.

Speaker A: Yeah. When did you first kind of start to see the AI's potential on what, what was possible?

Speaker D: I would say everybody knew the potential. I think it was the explosion of the models really. And what you could tangibly see when people were showing you, here's what we can do with it, and making it a lot easier to use because ultimately the user has to be able to understand it and be able to, to interrogate it.

Speaker A: And so Bruce, obviously, you know, being on the inside for so many years, you know, coming into the industry side, what, what were some of those things that stood out with, with Rhombus that, you know, made this like the, the right place for you and, and kind of like being able to solve that mission for the agency.

Speaker C: Well, when I, when I retired, I knew I wanted to go to work for a small company that was working in the intel space, or at least in the national security space and see if I could help them become a big company. It's Been a lot of fun. Um, the ability to pivot so quickly, to innovate as fast as we do. We work with partners, we identify really hard questions and relatively quickly, sometimes within days, we can provide them very, very solid answers. That's what I really enjoy. The speed at which we can operate in the private sector, the focus that we have on the mission that we need to get done. Uh, it's been a lot of fun.

Speaker A: Can you go a little bit deeper into, you know, how Ramis's Gen AI product does some of this in practice? I think you kind of alluded to the global view, but if you can go a little bit deeper, that'd be great.

Speaker C: Well, I'll give you some examples of things that we've done as a company. You know, in 2018, 2019 we were asked by um, uh, Diu and I, uh, think NATO special Operations if we could build a model to predict Taliban attacks at the provincial and district capital level in Afghanistan. Um, at the time, of course the entire community was, was uh, was getting smaller, but the intelligence requirement didn't get any smaller. So they needed every advantage they could get. And Dr. Roy, who's the CEO and founder of the company will tell you, he said at the time, there's no way you, you can't build a model like that. There's not enough data, there's not enough signal that you can pull from that to make a, a, a model that would be successful. But he learned, and the company learned, working with experts on the ground who really understood the Taliban, that you could, and we were actually quite successful, success in predicting those attacks, which allowed the US and partners to focus their intelligence collection on those areas where there was the highest probability of an attack happening. That ability to take an open source platform and lead directly to intelligence collection that's going to save the war fighter's life, uh, to me is phenomenal. And that is something that even 10 years ago I don't think we could have done. As Sheetal talked about how rapidly things have evolved, um, that ability to get after some of these hard problems. Another example that we did, it was around the same time frame we built a model, um, to map out the fentanyl trafficking networks using open source data, using dark web data and really successful model, um, and using AI not only to map it out to level, I've never seen that before, but then use what is the probability that a node, an individual is actually tied to narcotics trafficking. And AI was great at that. Again allows law enforcement then to focus their attention on those on those key players, um, sifting through tens of millions of data points, uh, instantly. So you can really get after some very hard, hard problems and produce results that the IC and the law enforcement can use immediately. And that to me, a thing I love about this company is that, you know, you join the IC because of the mission, you want to support the mission. You can still do that in the private sector, just in a different way. And you can drive innovation that feeds right back into the ic. And that's what I've enjoyed.

Speaker A: M. Sean, I'll let you take lead because I know you gotta.

Speaker B: Yeah, so already I think we're getting to some interesting territory here in terms of, um, you know, user needs, um, with AI for OSINT analysis or analysis of any kind of intelligence. Um, and like you both have said, things have changed rapidly since kind of the Cambrian, uh, explosion of LLMs and about the end of 2022, um, what's the future of the. Maybe not even the future, but the current state of the analyst profession. Right. Where for traditionally analysts come from linguistic backgrounds or geopolitical backgrounds, that kind of thing. But now they have access, um, not only to the raw data, but because of things like, like rhombus, um, sort of an intel, an intelligible level of information understanding because of AI and ML models behind that raw data or on top of that raw data. So how's that profession of doing the actual end user analysis change as this new technology is getting introduced into the ic? Uh, and I guess I'll, uh. Bruce, I think I'll start with you and then we'll go around the horn with this one.

Speaker C: Well, I'd rather start with Sheetal since definitely her area of expertise. I was a case officer, certainly happy to talk about it. So Sheetal, from your perspective,

Speaker D: just to start, I mean I haven't actually been an analyst in decades, but I would say the changes every day an analyst comes in, you look at your computer, you look through every single data stream and try and figure out, and you yourself, along with your colleagues is trying to figure out what is in that data and what came in overnight. Now you probably can get all that information, like here are the key things you need to look at and you have a little bit more time to actually think about what it means. Instead of trying to do all of that in a compressed timeline, you, uh, you know, the AI can go through all that data and you probably get a little bit more data. What I don't think it takes away from is an analyst's need to actually understand Culture and language inferences. So you still need to understand the country you're looking at because I don't think a computer is going to uh, at least not yet be able to give you some of those nuances. So I think it makes it easier frankly.

Speaker C: And actually situ, you just described what we're trying to do, which is how do we empower the analysts? How do we provide the analysts with different data sources they might not have before, or as you said, get them the information they need in a much faster way so they have time to do the assessment part. That's really, I think, what we want the analysts to do. What does it all mean? How do we empower that? Mhm. What's interesting from my perspective is to watch the evolution of tradecraft. Um, we were just having a conversation this morning about bias in AI. It's a question we frequently get. Can an AI model be biased? Of course the answer is yes. It's trained on human data, it's written and built by humans. There's always going to be bias and how you interpret it, there's bias, how you prompt it, um, could definitely lead to confirmation bias or the types of biases. So it's something that we're working hard on. How do you deal with that? Because we produce intelligence at the OSIC level and so we're developing new trade craft techniques. When you have a model saying there's an 80% probability that Russia is going to do something, how do you incorporate that into analysis? Um, so it's been a lot of fun to watch our analytic team kind of build out these capabilities. And this tradecraft, I think it's going

Speaker D: to get a little bit more complicated because you need to know the sources that the data is coming from and then weigh those sources. And if you don't understand the, the source of where it's coming from, then it becomes a little difficult. And that's where analysts will get a little, um, nervous about their own tradecraft because you do need to know the sourcing of it completely.

Speaker C: I just challenges with different commercial LLMs. You don't know the sourcing, you don't know how it's extracted, you don't know. Did they summarize the entire document? Did it summarize just part of the document? So I think for intelligence purposes and LLM has to be crafted by intelligence professionals, for intelligence professionals to make sure we're addressing what sheet will just talk about. Is the sourcing accurate? Can you use it? Any AI system that the black box to analyst to me is Worthless. Because if you don't know the sourcing and how it derived its response, how in the world do you use that?

Speaker A: Yeah, we're seeing that across a number of different verticals where it's like tailored AI plus, like we had a conversation with, on, um, you know, the soc. Right. Security. Security Operations center, and how it's used to empower security analysts versus replace them. Right. There's always going to need to be that kind of human in the loop element, but making sure that what's being built there is tailored to that specific craft. And I think that's kind of what you're alluding to, Bruce. I think that's fascinating. I'm curious about. Um, you know, we wouldn't be talking. We don't really have a conversation about national security or defense if we don't talk a little bit about some bottlenecks, uh, uh, to innovation. So I, I want to kind of just hone in on that a little bit and maybe just build a little bit of awareness for, you know, some of those areas where innovation stalls in the intel community. Because obviously, you know, data is something that is extremely secure and, uh, for good reason, uh, obviously need, need access to that data to, uh, effectively kind of train some of these different models. But let's hear it from the inside. You know, what. What were some of the things she thought that you saw were slowing adoption? You know, anything. Policy, people, process, you name it.

Speaker D: It's. A lot of it is actually, um. It's cultural. Some of it is cultural, some of it is procedural and just the bureaucratic nature of government procurement. And if you look at it, there are attempts to change a lot of this. Now, I think DoD is trying to figure out a way to do rapid acquisitions, but these small companies, if they don't have an in, or even if they have an in, if you don't understand the contracting vehicles or the way to go about doing it, if you don't have a partner that might already be accredited. Inside, it's all about partnering. And a really good story about how you're addressing mission, because you can get stuck. I mean, a lot of people say pilots are not the way to go, but I do think sometimes pilots are the way to go if you're trying to prove a concept.

Speaker A: Hm.

Speaker D: Because scaling immediately is not something that people are very comfortable with.

Speaker A: Yeah. Bruce, let's hear your side from the industry. What's the hardest part about introducing some of this new tech into the ic?

Speaker C: You know, I found, um. There's. There's very, there's buy in pretty quickly when we get to sit down with partners and say, okay, let's, let's look at your problem set. Let me show you how we can answer some of the intelligence questions that you have, you know, with our products. So there's enthusiasm, but then you run into what Sheetal talked about, which is the bureaucratic process, uh, of contracting. It's just, it's just a reality. And I agree there are, there are efforts underway to speed it up and some partners are better at it than others. Um, we're not a startup anymore, um, but I know early days, um, that really makes a difference for a small company if it takes a long time to get you on a contract. Uh, for our company, DIU was a fantastic partner. Really helped, um, to get the company established to the point where we are now. It's just a question of, I agree with Geeth a lot. Having those partnerships, having someone in the company who understands the contracting process. If, you know, once, once you've been in the business for a while and we've been around for 14 years now, you learn how to do it. But when you're an early stage company it's, it's all a mystery. So how do you, you, uh, know, how do you succeed? And you succeed by, by partnering with, with folks who know what they're doing.

Speaker A: Yeah, Sheetal, you, you stood up a mission center, you know, focused on emerging tech with Verus Sait. Perfect, uh, pairing here. Um, yeah, I mean, can you talk a little bit about that, like what you learned in that process and you know, maybe you can even pull a little bit, um, a little bit more on that thread around going from prototype to production and national security environments. Because a lot of the listeners on this show are early stage founders, a lot of them dual use, trying to figure out that commercial application, transitioning into these more regulated national security environments. Um, can you talk a little bit about some of those things that you've seen, uh, that were effective in going from prototype to production?

Speaker D: Um, sure, I can talk about it. I mean one of the things about our mission center is we weren't actually bringing technology into the building. We were trying to understand the ecosystem of the technologies and figure out who was doing what in those, uh, tech spaces. But as a result we did run across some amazing technology and tried to get it to the right people and make those introductions. But I think, uh, it still went back to the, if there is a way to bring it in. And there's a difference I think in the building and Bruce Correct me or add more to it. But if you're bringing in something for an operational purpose, it's a lot easier to bring it in for something like that. If you're bringing it in for an enterprise function that's going to be sitting on the infrastructure, that's a much longer, complicated process. And that's where sometimes things can get bogged down. But if you're a small startup and like Bruce was saying, and you're trying to scale, that might not necessarily be the way to scale if you're coming in to do an operational support task. So there is an inherent, um, battle, I would say when it comes to technology, we will utilize the technology for its purpose. But for a small startup, it might not work on the timelines that they are beholden to. Right. That's where DoD and others, which are much bigger entities with much bigger budgets, are better customers. The special operations community. There are things like that.

Speaker B: I think this highlights something that's really important, uh, in terms of a difference between commercial markets and mission in terms of defense and the ic, um, if you don't have a really firm mission purpose in mind, if you're a new technology company, you don't really have a deep understanding of what a customer's mission might be, it's really hard to get the time in front of them and frankly to build the trust for them to be interested in your technology if you're selling them what's sometimes artfully termed the art of the possible. All of these things you could do with this piece of technology. You'll get some meetings, you'll get motivated people listening to you. But if there's not a so what at the end of that, where it says, this is how I can really affect your mission in a positive way, it's hard to get that traction. But that can be in contrast to, uh, the approach to commercial markets. Commercial markets, you want to find a pain point with a customer, um, but they can also be interested in that are of the possible and things that I could do if I were to adopt this technology. Um, and I've seen that sometimes is a stumbling block for pure commercial firms with no background in the IC trying to work with IC partners because it's almost a different language, um, and how you present to folks. And so I'm wondering, um, again, question for everybody, but I'm wondering, um, Sheetal, if you've seen folks who've wanted to do work within the IC but who don't quite grasp that mission concept, and so they flounder a Little bit. Is that something you've seen before?

Speaker D: Yeah, we've seen it, uh, in a couple of instances. But I will also say there are people who actually address a mission need. But, but it's still hard, right? Because if you don't have a, uh, product to show, it's hard to get an organization to say, okay, I will invest in this with this idea that you have this product coming. Whereas if you have something and then you have your own manufacturing, your hardware and software stack, that's a lot easier for somebody to get behind.

Speaker B: Yeah. Bruce, I wonder if you've, if you have any additional kind of insight there. Again, because you've surfed both these worlds inside and outside of the building.

Speaker C: Well, inside the building I've actually had jobs where I was responsible for overseeing contracting, uh, for parts of the agency. And I can't agree more with Sheetal on that. You have to come in there with, hey, I understand your problem set. Here's a solution. We have a product, or at least we're close to a product, close enough that you can have confidence in us that we will deliver to get that attention. Because people in the IC are trying to solve hard problems every day and if it's the right solution, they'll take it. They don't often have the time to say, hey, let's do a pilot over time where you develop this out. And that can be hard for a small company, especially if they don't have experience in this space. You're not exactly sure what problem you're trying to solve, but you have a really fantastic technology that you developed. So trying to understand what is the problem that the client has. And if you're outside the ic, that could be hard. I do think, uh, working with DoD is a little bit easier, um, than it is with the IC for the reasons that she talked about. Bigger budget, a lot more potential customers. But m. Yeah, it can be tough.

Speaker A: Yeah. Bruce, I, I was uh, doing some research on Rhombus and something that kind of jumped out to me was the uh, Haven AI with the um, wildfire response into uh, proactive protection. Uh, so I guess detection, uh, within minutes of ignition of uh, wildfires, those types of scenarios. Right. I think it's always interesting when we think through companies, um, that have like a dual use approach, um, to their strategy with Rhombus, uh, specifically, I guess who was the uh, original kind of like customer set that you were uh, going for. And then when, when do you start to see these like, areas that uh, technology carries over into some of these Other industries. Because I think that's what's so fascinating of. You know, we see it a lot in uh, like law enforcement with like, you know, counter UAS technologies and stuff like that. But maybe just talk a little bit about that, uh, origin story and then how. Yeah, how, how uh, you, you've kind of seen your technology evolve into these other areas.

Speaker C: Yeah, it's a fascinating story. Um, in 2011, when the Fukushima Daiichi disaster happened, when, um, the need to clean up the plutonium from the nuclear reactor that was hit by the tsunami in Japan happened, Dr. Anchor, who's our CEO and founder, invented a neutron detector. He's got a PhD in chemical engineering from Michigan. Uh, he got a patent in Japan, won a contract to do the cleanup. He hired Dr. Sarah Cowan, our senior VP for ops and product. She has a material science PhD, and she built these. So he had done it on. She built them really successful contract in Japan to do the plutonium cleanup. Literally just by luck, Diux opened up right next door to the company in California and Anchu CEO and knocked on their door and said hi. That started 10 years of work in the, in the IC and uh, military space, taking the machine learning tools they developed to extract data from, from the neutron detectors, applying them to hard problems that the IC was trying to solve, whether it's, you know, fentanyl trafficking or terrorist attacks, and then the constant evolution. But we have another side of the business as well, where we actually help the Air Force, um, use AI for budgeting purposes, for logistics purposes. So we have really a breadth of things we cover, from budget and logistics to intelligence and predictive analytics. And it'll evolve over time. You really get to know a customer or a partner and say, these are your problem sets. Well, let's see how we can get after it. And maybe the problem is an intelligence, maybe it's logistics. Well, let's try to get after that. So that's kind of how the company has evolved over time.

Speaker A: Yes, that's fair. I love hearing those types of stories because it's um, you know, those, those use cases. You just, it's almost like, just creatively think through of like how that technology applies to this problem set. And you can quickly see it. Um, you referenced, you know, diu. Um, you know, we've had them on, uh, a few previous episodes. Love the work that stems out of diu. What are some of the other, I guess, um, outlets or channels that you would say, uh, could prove beneficial to, um, you know, earlier stage founders that are trying to navigate Some of those problem sets across defense or national security. Are there, uh, like, what advice, I guess, would you give to, you know, some of these founders that are just trying to. Yeah, yeah, again, like, navigate those waters? Um, DIU, obviously, is a great, uh, reference. Uh, but anything else that you've seen, um, through your progressions in your career or specifically out of, you know, Rhombus. Strategy.

Speaker C: Strategy, yeah. From the IC side. And it's not. Not from a rhombus issue, but something, um, that. That Sheetal and I were very involved in was, um, uh, managing the agency's relationship with In Q Tel. And of course, in Q Tel's phenomenal, uh, as an organization, they've had great successes over time. So there are capabilities like that out there, be it DIU or be it in Q Tel. But beyond that, it really comes down to what do you always need to be successful? Your network. How do you network? How do you develop the expertise that you need to understand what the problem set is? How do you get in front of people, um, to give them an opportunity to see your products and solutions? You can go the traditional route of responding to RFIs, and that can be successful, but for a small company, that can be quite hard, um, because there's real expertise in writing a successful, uh, uh, response to that. Uh, so I think networking is probably more important and understanding what you don't know and going out and getting the talent to help you answer those questions.

Speaker A: Yeah. Anything that you wanted to add on top of that?

Speaker D: Chito M. Uh, Bruce is absolutely right. And I think there are a couple of organizations that DoD also has. That's not just DIU, but I think Special Operations has its own entity. Air Force does. But there's also trade shows. That's where people go and look for a lot of the emerging technologies and what is out there and what's ready to be seen. I think some of it is just, as Bruce is saying, networking and finding those people that can be your champion.

Speaker A: I want to make sure that we also, uh, just kind of get a little bit of, um, a picture of Rhombus because, uh, maybe just some of the quick hits around the company in terms of size, um, funding. And where's the company headquartered? Where are you all kind of positioned around the globe?

Speaker C: So we're, uh. We're about 200 people. We're a bootstrap company. We have no external investment. Anjou, uh, took the revenue from our initial contract and has just built on top of that. Gives us a lot of freedom and flexibility, which I really Appreciate to pursue some of these really hard problems that that takes us a while to develop an answer to. Uh, we're headquartered in Palo Alto, but we have a large office in D.C. uh, to support our work with the Pentagon and the ic. We have an office in Japan, we have an office in India as uh, part of our uh, growing global program. Uh, so it's, it's, it's been fun to watch the company, even just the two years that I've been there, watch this grow and expand.

Speaker A: Very cool. And then I have to ask you to somebody that spent, yeah. You know, close to 30 years, uh, you know, within the IC, you know that I'm always curious just this is kind of like the recruiter, uh, in me as well, but the background or the uh, the transition to the private sector, you know, what, what uh, I guess what are some of those things that go through your mind in terms of, you know, knowing what you want or what you don't want when making that transition? Because it's, you know, you've been in an environment for so long. I'm always curious on, you know, what led you down the path of um, you know, consulting and you know, also getting involved in venture.

Speaker D: So I will say that, uh, when you're retiring and you feel like you have done everything you could do in your organization, it's a little liberating when you're trying to figure out what to do next. It's more of the working through the network, talking to everybody who has actually gone ahead of you. And their advice was really consistent. Figure out what you want to do. Do you want a full time job or do you want to do uh, a portfolio of different things? How do you want to, do you want to move? Do you not want to move? How do you want to give back? Do you want to be involved in the same kind of work or do you want to do different things? So for me it was pretty easy. I wanted to do a portfolio hodgepodge or a bunch of side hustles as I call it. I'm not ready to do full time. But the other thing was how can I take what I had learned and help in the commercial space without doing a full time job? So it is an interesting transition. Um, but it is also very different if you're coming out as a retiree versus somebody who is separating because they want, they already know what they want to do in some cases and they just, they launch ahead and then in a couple of years they change jobs. But I, I will say our network is Pretty good. Nobody says no to a conversation.

Speaker B: Yeah, so let me, let me piggyback off that cheetah with asking that question, uh, at the beginning as opposed to the end. So, um, in terms of a career, um, if you're a technologist in the commercial world, there's kind of a few different established pathways for what a career looks like. You can be very hands on as an engineer, you can go into product management, you can go into general leadership, investment, what have you. Um, in Federal Service, I think it's actually still way less defined about what a good career in technology inside of the ic, let's say, um, looks like because people think of the IC and let's be honest, they think about James Bond and other things like that at first, maybe if they're younger, if I'm a computer science major coming out of a, uh, university and I have this urge for service, but I don't know exactly how to plug in what's the beginning or even up to a mid range, you know, middle of a career look like for someone who wants to do technology, but in that, that service, um, orientation, say within the ic.

Speaker D: So in the ic, I mean I will speak to CIA. I don't know how they come in in dia, but at uh, CIA, if you're coming in your first several years, regardless of which of the five directorates you're coming in, you're going to be training, you're going to be learning the tradecraft of your directorate, you're going to be learning the expertise you need to work at CIA, figuring out the building, figuring out how to do, getting all your logins, but also passing all the training you need. And whether you're computer science, you're coming in as a demographer, you're coming in as a case officer as boosted, there are set paths for at least the first several years.

Speaker B: Okay.

Speaker D: Uh, once you're established as an expert, then people's careers kind of vary. I started off as an analyst, but ended up going overseas and as a deputy chief of station. And so you can have unique paths, but I would say when you come in, learn your craft really well for the first couple of years.

Speaker A: Yeah, I'd love to hear Bruce's, uh, uh, perspective too. And then I kind of want to ask that same question because Bruce, you, you went from, you know, 30 years, retired, and then a fast paced startup kind of m, you know, culture. Uh, you know, let's, let's hear a little bit about that and how, how that shift kind uh, of went for you and maybe what Your expectations were. And then what, you know, what it really was.

Speaker C: So Sheetla talked about, um, you know, when he stood up T2MC, we got to meet so many companies, uh, doing really remarkable things. And I found that very exciting. I mean, back many, many decades ago, before I joined the government, I was a math teacher. So I've always been interested in mathematics and science. And I saw what companies were doing and I was like, I want to be part of that. So as Shaitl said, when you retire, you have to make decision. What do you want to do? Well, I knew what I wanted to do. I wanted to work for a small company in the national security space, hopefully one using, uh, AI, because I just find it personally interesting. And I was very fortunate to find Rhombus, um, and come on board with them not long after I retired, which it was exactly what I wanted to do. And then help build out our intelligence program. And it's been a lot of fun. Um, it's been hard. I can tell you without a doubt, the hours that I work, um, surpass most, not my entire career. There are parts of my career inside the government. I work a lot more hours, but on the whole it is. But, uh, that's to be expected. You know, you're growing a company and there's a lot to be done there. Uh, but it's a very supportive group and it's, the work is fun. It's exactly what I was hoping for when I retired.

Speaker A: Yeah, that's great that you're able to kind of get those two different perspectives. And that's why I was excited to have you on the show because, you know, we oftentimes will have, you know, those, uh, you know, those folks that come from, you know, years in service, and then maybe they start a military tech startup or what have you. And it's always intriguing to kind of catch, you know, what it was, uh, like making that transition and kind of building a little bit more of that for our audience. Um, I think, uh, we can kind of put a bow on the main discussion and transition, uh, us into the final segment here, which is just a fun, rapid fire Q and A segment. We call it the five second scramble. Um, you know, just quick, quick answers, gut answers. Try, uh, to keep it within five seconds. Uh, Sean, maybe you lead us off with, uh, Sheetal, and then I'll finish, uh, with Bruce.

Speaker B: Yeah, absolutely. All right, Chili, you ready? Again, it's Quick Answers.

Speaker A: All right.

Speaker D: Do we get the jeopardy?

Speaker A: Oh, uh, man, I can tee that up. Let me see if I, you know,

Speaker B: we don't have the buzzer ready, but maybe for future episodes. No one's ever come up with that idea before.

Speaker A: I'll get a sound bite for the next one.

Speaker B: Yeah, and usually how I do these, my style of doing these, Sheetal, is they start off a little bit, maybe more trickier, but get easier as we. As we progress through. But don't worry, nothing's going to be cataclysmic here. Um, we were just talking about your transition from. From the IC into commercial. Uh, what would be the single biggest adjustment you think you've had to make with that within that transition?

Speaker D: Learning the private sector in a way before.

Speaker B: Awesome. Very different now. Uh, as part of that transition, too, there's different cultural kind of artifacts. Um, what's one part of the culture at CIA that you wanted to maintain and preserve as you transition, you know, into commercial?

Speaker D: The mission focus. The.

Speaker B: That's a. That's a common one.

Speaker D: We hear national security.

Speaker B: Yeah, yeah, super common one. And that's, ah, a, uh, I'm always happy when, when our guests say that. Um, all right, so what's the best leadership advice you've ever received?

Speaker D: Don't take yourself too seriously.

Speaker B: Classic. Classic. Um, what's the most surprising skill that you've had to learn in your career?

Speaker D: Ooh. Ah, I have to think about this.

Speaker B: All right, good.

Speaker D: I've learned a lot.

Speaker B: Is there any skill that stands out as odd that you weren't anticipating that you would need to learn or, uh, anything that turned into a big deal that when you first learned it, maybe you thought, oh, I'll never really use this. Something like that.

Speaker D: I would say probably some of my training that I got before going overseas into war zones.

Speaker B: Um, if you weren't. All right, so you've made it through. Those are the trickier ones. So now we're getting a little bit. We'll have a little bit more fun with these. But if you weren't in national security, if you weren't intelligence, if you're in totally different occupation, what do you think you'd be doing?

Speaker D: Probably a kindergarten teacher.

Speaker A: Really?

Speaker D: Yeah. Five year olds just make me smile. I think they'd be a lot of fun.

Speaker B: I think those would be some pretty. Pretty with it.

Speaker D: Five year olds, you know, sets that you need. Right. To manage them.

Speaker A: You, um, either love that job or you hate that job. I feel like,

Speaker B: uh, awesome. That one. Unexpected. Very cool. Um, what's a trend right now with AI that everyone's excited about that you think you're more skeptical about or Vice versa. Things that maybe the industry is skeptical about, that you're more, um, bullish on.

Speaker D: I'm not really bullish on AI. I am a big fan of AI. I'm a big fan of, uh, the models. I think what I'm a little bullish on is the blind acceptance of whatever sometimes you get from. From your questions. Because if your question and your prompt is not precise, you're going to get something back that might not necessarily be accurate. And if you don't know it's not accurate, you're not going to know what you don't know.

Speaker B: I. So I really, really like that response. I think this is one of.

Speaker A: The.

Speaker B: One of the clips that Tim and I will make later is this retired CIA technologist is telling people to be a lot more suspicious of AI Outstanding advice. I think it's very, very.

Speaker D: Technologist.

Speaker A: That's right. Uh, then turned kindergarten teacher. Like, this is the plot of a movie for sure.

Speaker D: Like the Kindergarten Cop.

Speaker A: The Kindergarten Cop, that's right.

Speaker B: Um, all right. Uh, this one I've been dying to ask you. What's the strangest place you've ever had to take an important work call?

Speaker D: Oh, A, uh, public restroom in an overseas country.

Speaker B: I totally understand if that's as detailed as we can get, but. Awesome answer.

Speaker A: The rest is classified.

Speaker B: Exactly. The rest. Yeah, we got to classify it. Um, and then your best purchase under $50 under. In the past year.

Speaker D: Oh, in the past year. So I finally bought a laptop because I didn't have one, so it would be my case for my laptop because I tend to drop it everywhere.

Speaker C: Awesome.

Speaker B: Um, and then we always end with our, uh, kind of traditional question, which is, is there a corporate philanthropic organization or a charity, something like that, that is near and dear to your heart and that you want, um, our listeners to be aware of?

Speaker D: Oh, yeah. I sit on the board of my dad's nonprofit. It's called the SAVAK Project, and it is providing rural health care to villagers in India. And it's based on. He's a retired Navy captain, so he's based it on the naval corpsman, the, um, medics in the Navy. And it's basically training villagers to provide basic health care. So it's saveupproject.org and that's s E

Speaker B: V A K. Do I have that right?

Speaker D: Yes. Yeah. All right.

Speaker B: Civicproduct.org Run out.

Speaker A: Very cool.

Speaker B: As all good charities are sheetal. Uh, you nailed it. Thanks so much for doing the five second scramble. And, yeah, we'll pass it over to Tim And Bruce.

Speaker A: All right, Bruce. Ready?

Speaker C: I want the Jeopardy. Music.

Speaker D: Yeah.

Speaker A: It seems like a fan favorite. We gotta queue it up. I did have a soundbite previously that was just like a eggs dropping into a skillet with that sizzle. That was, that was the sound bite we had. But I kind of like the Jeopardy. Idea. Um, all right, here we go. Uh, if Rhombus Power were a band, what genre would you play?

Speaker C: I don't even want to begin to answer that question.

Speaker D: Somebody's gonna get offended.

Speaker C: Yeah, exactly. Right. I think it depends on what the partner wants.

Speaker B: That means that they're like a wedding band. Right. Where good wedding bands can kind of, can kind of play any tune.

Speaker C: Uh, what our client needs, we'll play.

Speaker A: That's right. Okay. I like it. Uh, we'll keep an open mind on it. What are the most critical roles that you, uh, guys are hiring for over the next three to six months?

Speaker C: Um, right now we're actually hiring for a senior Director of intelligence to work with me as we expand our intel program. It's a, it's going to be a great job. And because we're expanding very much in the geospatial area and also in traditional analysis, we're looking for people with expertise in both sides. We're always looking for talent. Uh, on the development side, um, it's a great place to work. So it just depends on what your background and interests are. But there's probably a role for you at Rhombus. Ah, in one of those spaces.

Speaker A: Very cool. What's something about the culture at Rhombus that you would want new employees to know about?

Speaker C: We have fun as a company. Um, we enjoy the work. It's challenging, but it's a great environment.

Speaker A: Very cool. What's the most surprising skill that you've had to learn in your career?

Speaker C: How to manage and execute federal budgets. That was a 90 degree turn that was not expected. Coming out from being a chief of station overseas and coming back to Washington, my boss saying, I really want you to manage the budget process. Really fabulous job. It really opened up many doors for me. But federal budgeting is not, is not intuitive. At least it wasn't to me. So it took a while.

Speaker A: I'd say that's the number one answer for a lot of folks, probably. All right, finish this sentence. AI and the IC will succeed if blank.

Speaker C: If people are more open minded about what the potential that it can bring and willing to develop new trade craft.

Speaker A: Very cool. Nice answer. Favorite, uh, travel destination for personal travel.

Speaker C: Um, anywhere in Oaxaca, Mexico.

Speaker A: Nice what was your very first job

Speaker C: right out of college? I was teaching mathematics.

Speaker A: Oh, wow. Very different from my. I worked at Domino's.

Speaker C: Ah. Doing my first job when I was.

Speaker A: Yeah. What's your very first. With the very first job, like, I

Speaker C: worked in a, um, stand at the baseball park, selling hot dogs and popcorn, things like that. So network industry in Kentucky, I've done a whole lot of different things.

Speaker A: Very cool. So, uh, if you had to teach a master class on something that has nothing to do with your job, what, what would it be?

Speaker C: Uh, it would. How to. How to network, how to build connections, how to communicate effectively. Because that's fundamental to. I don't care what, what your life is, what your career is. If you don't know how to do that, you're going to struggle.

Speaker A: Yeah. Well said. Uh, favorite CIA or Secret Service related movie?

Speaker C: Uh, Slow Horses. It's not a movie, but.

Speaker D: Oh my God, that's amazing. Yes.

Speaker C: Love that show because it doesn't take itself seriously.

Speaker A: Uh, and then last one, uh, what's a charity or corporate philanthropy that's near and dear to you?

Speaker C: So my sister worked for the Red Cross for many years, so I've always been a huge fan of the Red Cross.

Speaker A: Very cool. Yeah, we'll get both those, uh, a plug in, uh, the show notes when we release the episode. And, uh, just wanted to thank you both for your time. Thanks, uh, for sharing your insights on modernizing the IC and obviously the very cool technology coming out of Rhombus. Um, really appreciate you all joining us on the podcast.

Speaker D: Thank you very much. This is fun.

Speaker A: Yeah. Thank you guys.