E91: Forward Deployed Engineers: The 170-Million-Job Earthquake That Will Redefine AI Work Forever

(upbeat music) Okay, so you know what's absolutely fascinating? So there are these farmers out in Iowa and Nebraska, somewhere in the American Midwest, and they're spraying chemicals over their entire fields. I mean, everybody does, right? And probably just like Americans do it, you know, just blast everything, okay? Chemicals everywhere, never ending. And this year, this year was the first year that they reduced their chemical usage. Now you would say, okay, by what? 2%, 5%, between 60 to 70%. Now you might think, okay, but if they're using so fewer chemicals, they must have a problem with their yields, with their crops, right? Less input, less output, right? Or maybe they're doing it for environmental reasons, who knows, right? Well, no. First of all, less chemicals. Secondly, better yield. How are they doing that? Well, they're spraying more, but precisely more targeted with less waste. And you might be thinking, okay, cool Malcolm, that's like some new agricultural technology GPS system, some drones, kind of a bit, yes or no, you know? And you know, this podcast is about AI and the technology that made this possible was models from OpenAI. Yes, chat GPT. But here's the thing that's really revolutionary. The technology didn't just magically appear. John Deere didn't just go and download chat GPT and say, oh, what do we need to do? No, they called somebody from OpenAI and they said, could you please send some people over to look at our processes and figure out how we can improve this word AI? Isn't this wild? And these people that they sent over, they have a very specific job title. It's a job title which really didn't even exist one year ago, but this entire episode is about this job. It's the fastest growing job in the entire AI industry right now. And these people are called forward deployment engineers. And that's what we're gonna talk about today. So yeah, action, let's go. ♪ Flip the switch, the circuits live ♪ ♪ Digital dreams, they thrive and strive ♪ ♪ Press a piece, code it to feature mine ♪ ♪ Serving knowledge of the AI kind ♪ ♪ Welcome to that AI cookbook show ♪ ♪ With the data streams and the insights flow ♪ ♪ Shoot and slow, let the rhythm grow ♪ ♪ For every soul in the know and those who don't ♪ So, our forward deployment engineers, we're gonna call them FDEs. Now, forward deployed, it's actually a military language, so I work sometimes with CEOs of huge companies and also smaller ones, and some of these CEOs, believe it or not, I mean, they're CEOs, but they still have time to be active in the military today. And they would love this not only this episode, but this understanding. Because in military speak, forward deployed means send them first. Let them run ahead, let them scout what's happening before we even send anybody in, or for example, before we even send any tanks. So a tank grenade here, that's a position that you send out and they need to like, you know, break down everything that they can, you know, and they do this before the entire battalion arrives. So, but this forward deployment engineer wisest our military thing, well, forward deployment engineers, they are at the frontline. They actually get deployed in the field. So if you look at, for example, John Deus' tractors, they probably send these forward deployment engineers on the factory floor, where they manufacture. They send them out in the fields, where they are the crops. They send them to the guys who are repairing the tractors. So for example, if you have a tank, you would send it to an operations center. And they even send these forward deployment engineers to military bases in Afghanistan. Military bases in Afghanistan? Yes, because this is actually not such a new job. This role forward deployment engineer was actually developed in the mid 2000s. It was developed by a company that everybody loves to hate, which is called Palantir. And Palantir, they're kind of these big data analyst firms. And they were founded in 2003. And from the earliest days, they literally borrowed this concept from the military. So soldiers who get forward deployed to foreign countries. And Palantir said, well, if soldiers are doing that, if the military is doing that, why don't we deploy our software engineers in the same way? And they started sending their software engineers not to conference rooms in fancy offices like in Behrner, Bazelotsewich, but to Africa military bases, Iraqi military bases, factory flows in the American Midwest, real places where real work happens and real problems need to be solved. And if you look at their system, it's absolutely brilliant. Now, how do software companies work now? Well, they will send a project manager. They will send an account manager. They will send a guy like me, Malcolm Goals, he talks to customers, he tries to understand what they want. And then you go back and you talk to the developers. You say, look, this is what the customer needs. And in this moment alone, a hell of a lot of information already gets lost in that step. Now, apparently they said, okay, let's create two different customers. Echo and Delta. Now, what does Echo's job is? Echo is, they go and figure out what the customer actually wants, not what the customer says they would like to have. It's like no. Let's sit down and actually find out what your problems are. So they kind of let the translator to the customer. And Delta's job, they build it right there, on-site while the customer watches. And they work in teams, you know, they have multiple Echo and Delta teams working together like a platoon that was deployed to solve a specific problem. And apparently, I've been doing this for 15 years now. At that peak, FDE is made up more than half of the entire workforce. 50% of Palantir's workforce were forward deployment engineers. And think about it. Hey, the software companies you guys are working with. Have you ever seen a developer on-site? And if you did, they came over, they said hi and they left. Or maybe they did the handover for like two days and then they left. And then they disappear back into their home. Now, imagine a software company comes to you and says, you know what? We build better solutions faster. We understand our customers better. How do we do that? Well, we need an office where we can put four, five software engineers and we don't build software and hide in our cave. We bring the engineers, we understand what your problems are. And we sit in your office and we build your software right here for you. That's the Palantir model. So, Palantir, they built the system quite some time ago. And now, all of these AI companies that really have only come into existence. The last two, three years are adopting it. Open AI and drop it, et cetera. All of them that were born in the last three years are saying, we've given people to our GPT and they're not using it. Or they're just using it a bit. How could we fundamentally change the way that they work to help them? Well, we need to send people on the ground. And now, if you go to Indeed, you know, everybody thinks LinkedIn is such a big job platform. The biggest job platform in the world is Indeed. Because Indeed is an aggregator, you know. And if you look on Indeed, the monthly job postings for forward deployment engineers between January 2025 and September 2025, in just nine months, there has been an increase of 800% in FDE job postings. In nine months, which means 800% is not to trend anymore. It's not like it's like an explosion. It's a new job category being born in real time. Now, you might say, I don't get it, man, because we also have developers. I mean, you know, we have even developers in house. But this is where I think we would be totally wrong to think that way. Now, why does it matter? Because AI alone isn't enough. So Arno Founier, he leads OpenAI's FDE team in New York and in the Middle East. So he talked to the financial times. And he said, well, OpenAI expects to expand to about 50 engineers in this FDE team by year end. And both OpenAI and the tropics in the same, they say, the team that is scaling the most is the FDE team. Why not because they want? Because demand has quintupled. Five times more customers are coming to them and saying, OK, you've given us charity, you've given us the API. But please come on site and help us. Because now companies go and give their employees a year of AI training or Microsoft co-pilot or whatever, and people don't use the shed. Let's be honest, billions and billions and billions is poured into AI training, which is awesome for me and my business model. Companies are calling me and saying, can you train my employees? But to do a real business impact and we see that with our customers, we call them the builders. It's not people who give AI to the employees. It's people who need to understand the technology. OK, this is AI, this is charity. And need to understand the actual business. And for this, you don't need 100% pure developer. You need somebody who can combine coding or building with talking to people. And this hybrid role is really the future of work. And there's another great article I hope you can read it. It's written by Naomi Alderman on "Substack." She's a writer, you know, really, really beautiful stuff. And she said something very, when I would be poetic now, no shallow, but very profound. She says, look, AI is nothing special. It's not, it's just a technology. It's not a God, it's not a monster. It's like the Internet. But she says, but you know what? These AI models, they're trained on billions of parameters, which means, while they're not anything special, wait, they understand very well how our back works, they understand very well precision agriculture by John Deere. And you still need the human judgment. You still need that human in the loop that's human in the middle. I don't want to be too theoretical in these podcasts. Let's talk about a forward deployment engineer who we have at our firm. So his name is Damien. Now, what makes Damien's role so critical? So when we want to try to help companies with AI, what do we do? We do interviews. We cannot say, what's your job, what's hard about your job? What do you hate about your job? By the way, probably the best question you can ask in a company to identify AI use cases. Because people normally hate mundane stuff, the hate stuff that brings little value, etc. And then they would maybe say something like, okay, I get data from SAP. I have to upload it, I have to merge it, I have to put it in Excel, etc. And they say, ABCD, EFG. And we are thinking, oh my God, that's probably relatively easy for an AI. And who does these customer interviews? Now it's not Marsha, it's not to me, it's not porn, it's not Alex, it's not Malcolm. Damien does them. But wait, man, I thought you said Damien is a developer, yeah. But Damien does the customer interviews so that he can really understand what customers moved. Because when you're running a software house, normally you hire middlemen, so who are these middlemen? Well, he account managers, sales managers, you know, they talk to customers and they look good and everybody thinks everything's awesome. But for AI rollout, cut the middleman, you don't need these middlemen, the key account managers, sales managers, etc. Put them directly in contact with the developers from day one. And Damien, he doesn't talk to other software engineers. Damien does the interviews with CEOs of companies, CTOs, head of AI's, but also a secretary or somebody who's working in HR so that he's really able to understand their pain points. And that's really not technically impressive, that's really actually easy to do. It's humanly impressive. Okay, Martin, you got me. But how do I find these forward deployment engineers? Well, the thing is, mostly the people who listen to this podcast, at least based on our Spotify stats, which is not a podcast or the other platforms, but they're mostly between 45 to 55 years old. Do you know that you might be the perfect forward deployed engineer? Huh? I thought I need to quote. No. Well, if you listen to this episode, what did we say? FTEs are mostly young programmers, right? No. FTEs are mostly the most senior people that you have. They're often managers with gray hair. Why? Because it's easier and I say gray hair respectfully. You know, these are people who have learned for 20 years. This is the business concept, context, that's our company works. These are the people who have normally strong communications strength, you know? They are able to talk to sea level and in next minute to construction worker. And they have normally relatively good judgment, like an AI solution, they will understand, this makes sense, or it's just a word of salad, word of salad, you guys have this expression. Now, and they're very good at stakeholder management. So essentially, it's easier to bring a business leader by how to set up a console, play what Claude called, try cursor, try replete and learn a bit of Python than it is to bring 20 years of experience to a developer. Because with these basic technical competencies, you can talk to developers. You can maybe build a quick prototype and say, look guys, that's more or less, I think, for the customer once. How do I know? I did it. And I also already showed it to the customer and they validated it. And therefore, it requires people who can work at that interface. You know, kind of the interface where echo and delta are working together. And now, let's trash my favorite category of managers, HR managers, sorry, but sorry, you guys are all real AI in the end of the house. And I'll tell you that straight, because think a bit about it. So when we do workshops with HR people, one of the first questions we asked is, are you hiring people who know how to use AI? You know, but when you look at the job description and, you know, I had not now, but earlier, I had five job descriptions of the top five companies in Austria because we want to do a TikTok channel, well in Germany, what's written on the qualifications Microsoft Office? Guys, wait, my dad is 75. He can use Microsoft Office. Why are you putting on your job description Microsoft Office? This technology exists since 30 years. Why don't you put, have you used co-pilot before? What's your expertise in charge GPT and drop it? Have you used replete? Can you build prototypes with AI? Because if you hire somebody who brings AI into your new role, this is the biggest ever you could have in every single company in the world right now. Some people who actually understand AI, not developers, business people and on paper, those are forward deployed engineers. So my friends in HR go out there and hire your forward deployed engineers. But the first question when you interview them should not be, can you use Microsoft Office is what are you doing with AI? Is those are the skills that are needed? Okay, finally, now if you are mid-career and you're worried AI will take your job, please focus on developing skills that AI cannot replicate, the stakeholder management that you're going to do that, the talking to people, the project management, all of these things you have learned how to do the last 20 years. And if you've never used Python or replete, go and learn this. On the technical side you don't have a lot of work. Now remember, this is a job description that has grown on indeed by 800% in the last three quarters of a year. So yeah, that was it, greetings from Brigands today and I hope you enjoyed forward deployed engineers. Arya, mark them out.