BUILD

SUCCEED

David Chen: [00:00:00] no matter how good the underlying tech is, if you present the user that's using this as their livelihood and it, something breaks or doesn't work or costs them an extra hour to figure out how to do something, They're just going to throw it away and never use it again.

 Hi David, thanks for coming on Build to Succeed.

Hey, David. Great to see you again. A pleasure.

David SP-VG: let's get started with, why don't you give us an introduction.

David Chen: Yeah, I'm David Chen. I'm the head of software services at a company called Agtonomy. We're building autonomous tractors to work in permanent crops, and our mission is to solve the massive labor [00:01:00] shortage that's happening in ag today.

David SP-VG: Autonomous, right? So I was just out in San Francisco, actually saw you a few weeks ago, and, got my, first experience in a Waymo. So you're at the,in the center of autonomy and autonomous vehicles right now. what's the vibe out there with the whole autonomous movement?

David Chen: it's incredible. You'll see Waymo is driving all over the place. And after becoming a user, you realize you never want to ride in a human operated Uber or Lyft ever again, because they just drive so much better. It's comfortable. You don't get whiplash from terrible drivers and. You don't get your ride cancelled because there's, traffic.

can't wait till it's everywhere.

David SP-VG: It was, uh, eye opening for me, personally, cause the cars were both, more aggressive than I expected them to be somehow. And, interestingly enough, I feel like there was one situation where a car pulled over. And it was kind of still in our lane a little bit where even as a human, I was like, are we clear?

But it's interesting because it had [00:02:00] so much data that it knew it was clear. And we were all like, Ooh, that's tight, you know, but it's still pushed right through. And, um, so that was, that was pretty cool. And,we had a joke because we were, we got, we took a regular Uber and we were just talking about work, obviously what we do, we build apps and stuff.

And the Uber drivers like started pitching us his app idea, which is not something that happens in a way, Mo. So, you know, it's interesting. Um, but, yeah, so your career has spent a lot of time thinking about autonomy, right? And, I'd love to just tell us a little bit about your story arc. How did you end up at Agtonomy?

David Chen: Sure. So, growing up, I was exposed to computers very early on. Got into programming, learned how to program. And one of my first jobs, this was in the 8th grade, was Building a database program for a local, local company that made inflatable boats. And then from there, got other jobs, built doing PHP, web development, and that's kind of what led me down this path.

Went to school, did computer [00:03:00] science. Did my own startups after that. Eventually I worked at Twitter for a number of years when they were still much smaller. and it was around that time I got really into drones. this was a new thing, right? Quadcopters didn't really exist for the public. At least it was just something that can hover in place.

You can fly it easily, control it through an app. And that really sucked me in. I got into drone racing, building drones. And that's what led me to autonomous drones, and I joined a company called Skycatch, who was building autonomous drone systems to map sites like construction and mining. And it was just really interesting work on what, what this technology really enabled. That led me down to other robotics interests and that ultimately led me to economy where we're now building robots for farming, and bringing autonomous technology to help with labor shortage where people just can't [00:04:00] hire enough.

David SP-VG: Wow. I mean, that's an awesome set of experiences and you can kind of connect the dots there. You mentioned about Skycatch and the drones kind of doing like mapping. Can you get,what were you guys doing there? Were you making like 3D maps of these sites?

David Chen: Yeah, exactly. So for mining, it was open pit mines. Essentially, they dig a giant hole in the ground to extract minerals. Um, and it's constantly changing as they blast and remove materials for processing. They need to know exactly where things are and how much volume of material has been moved.

Traditionally, a human goes out there with a GNSS rover and they're measuring point by point and then interpolating that to create a surface. But it's dangerous because you can imagine right after you blast an area, it's unstable or you can't measure to the very edge of a cliff [00:05:00] because it's super dangerous.

So you're kind of measuring couple feet behind and then just projecting that point to give an estimate. So it's not very accurate. What drones allowed us to do was reverse the process where instead of human going out collecting data point by point, you can send a drone out, map the entire area.

Generate 3D terrain that's accurate to a few centimeters, and then you can do all the measurements you need on a computer. So that takes a human out of very dangerous environments, it makes the work faster, and getting that data was essential to the operations of these mines, and the surveyor who goes out there to measure it.

It's kind of that bottleneck. Everybody's asking for data. Can you go measure this? Measure that? Calculate these things. So it's on them. And these mines are in very remote, often very hostile environments. So being out there, just going outdoors in some of these areas could be very tough [00:06:00] and especially in polar regions where temperatures are, negative throughout the year and you get maybe a couple hours of sunlight.

Every second count and being able to do your work efficiently is all that matters.

David SP-VG: Hmm. Wow. That's, so I think a lot of people when they think of autonomous, they think of autonomous car where you're replacing a human driver and there's some concerns about jobs and things like that. Right. And it's also interesting 'cause you guys that autonomy are solving that problem 'cause there aren't enough people to do the job.

But in this case it's like autonomy plays a critical role. where it's just dangerous for a human operator and you're not able to even get the level of data, the quality of data you really need. And so the autonomy really helps you in that regard. These sound like really crazy places to go. Like as a, as an engineer developing software, I would imagine it would be quite difficult to like, Test the things you're doing given that like a drone has to be in a remote mining place like Flying around.

How did [00:07:00] you kind of like overcome some of the sort of like actual use cases of these things and learn and improve your products?

David Chen: Well, what we found is to build products for certain markets, it's not enough just to think about and talk to the people who are going to use it. You really need to live in their shoes to experience what they experience on a day to day basis. And for us, that means actually going out to the mine sites, working with them, helping them, training them to use these systems, and then actually experiencing what that process is like.

So for me, I spent some time up in the Arctic Circle at, A place called Red Dog Mine. It's a super remote mine, but I think a large amount of nickel and lead is mined from this single site that supplies the entire world. So it's very critical for all sorts of, production downstream of that. flying up there, you take a Alaskan [00:08:00] airline, it's a special charter that flies to the mine twice a week, they have their own power plant, they have their own runway, essentially you fly in, there's a compound where a couple hundred people live and work and they go on these crazy schedules of two weeks on site, two weeks off site, pretty much working nonstop in like these insane environments just to mine this material.

And. Every minute of downtime is, costs them money. So they need products that just work in any kind of environment.

David SP-VG: Wow, and so what are like when you're there and you're in the Arctic Circle And this product you're working on probably normally from your, your office, right? Committing code, what kind of things were you able to do and learn on site, as part of that product development experience?

David Chen: Yeah, sure. So the things that you don't think about when you're just developing software, because for software you think about, Oh, uh, you know, it loads, it does the right things. It, but in [00:09:00] these environments you have to think about. It's negative 20 outside. So what does it take to open up a kit, take out the drone, set it up, get it ready to fly?

Because that actually is challenging because you're wearing gloves. So we wanted to design systems where it was really easy to set it up. You essentially have everything ready to go. You pull it out, power it on, and it's able to do its job. and getting that data is also important. Where do you process that data?

In normal world, you can say, all right, just go upload this 10 gigabytes of photos to the cloud, get it processed. But at that mind, their only link to the internet was satellite uplink. So it was very limited bandwidth, which made that completely impossible. So what made our systems work is that we develop solutions to process this data on the edge in the field.

So by the time they're done flying, they can Load that data from the drone into our edge [00:10:00] compute device. And by the time they've driven that 30 minute ride from the bottom of the pit up back to their office, data's ready to use. They can pop it into their computers and get on with the next step.

David SP-VG: Wow, that's awesome. Uh, yeah, it's definitely a more rounded way of thinking about your products. Like, you gotta get out there and it influences your The actual software you write and solving some of those complex engineering challenges. So how'd you end up from mining then into ag tech?

David Chen: yeah, so I was looking for a change. I'd been at. Skycatch for almost eight years. And a former colleague was at a company called Ektonomy and they were building autonomous tractors. I went out to, uh, founder has a vineyard up in Sonoma. I went to visit them, saw the tractors in action, and it was really exciting.

And. Seeing it do work on its own and being able to be remotely controlled and autonomously drive around in a [00:11:00] lot.and for me, it was always finding something that not only was I interested in working on, but also what good does it bring to the world if I contribute my time and energy to making this successful product.

You know, that's something that drove every single decision that I made in life is like, this is going to have a bigger impact.

David SP-VG: Was there anything specific around these like hardware, you know, both, of these, um, career experiences you're describing have very heavy hardware angles to what they're doing. And you can contrast that to a lot of software engineers who are building kind of pure software. Like if you're working at Facebook or something, like your entire, Product lives as software, essentially.

You're, building software that integrates with, real things that are flying around or driving around. Was that, purposeful or was that luck or is there something that just really excites you about, the hardware software connection?

David Chen: I think that was purposeful. you know, ever since I was young, I would love [00:12:00] building things. that's always been something I've been interested in. In the early part of my career, I did pure software, building web apps and mobile apps. But once I got sucked into drones, it re revived that passion for robotics that I had as a younger and,led me to focus on career choices that involved hardware, because I just think it's so much more interesting when the software that you're building actually interacts with the real world.

David SP-VG: and the humans in the world, too, right? But that's interesting, because you have, these autonomous vehicles you're creating Uh, the drones or the tractors, but one of the key things you were describing is still the human element of like, I gotta open this box and put it out there.

so that's, I just think that's an interesting thing where it's this human hardware, software interaction. I think a lot of people out there maybe don't have a lot of that experience. Is there unique aspects of, as a software engineer, that you think are fundamentally different when you're developing for this sort of hardware, [00:13:00] these interactions like this, as opposed to just a pure, Map or website or something like that.

David Chen: Um, yeah, I think there is some differences. But I think it all comes down to understanding what the user is trying to do. I think a lot of software is built technology first, and then you give someone a product that you come up with. Whereas the products that I've worked on in this robotics realm, it's first understanding what is the traditional workflow that they're doing, What is their steps?

And then instead of trying to give them something new, you bring them something that's familiar, but is doing their original task in a much easier or much more streamlined way. So really understanding, their day to day work is extremely important.

David SP-VG: so you you worked in two very specific Companies [00:14:00] right and I'm curious if there were any sort of key tech decisions You made at those companies that really stand out or any sort of even I'd love to know also even Commonalities between those two companies or lessons learned maybe a sky catch if we go back there were there things that like when you were doing that when you're working on that and all the Environmental concerns you were thinking about and all the Product Concerns.

I mean, one of the things we're really looking for is how do you optimize success in these efforts? You know, we have limited time, limited budget. What were some tech decisions you made that you think helped you to be successful as a, both as an engineer maybe and as a company back then?

David Chen: Yeah. I think, for Skycatch. When we were creating the company, there's a lot of hype around drones. A lot of companies building drone software. what we really honed in on is what is the critical piece of this whole system that's really important. And what we found is, The core technology that does the 3D reconstruction was a core thing that we [00:15:00] decided we must own.

I was able to hire a computer vision team to actually build our own proprietary 3D reconstruction engine and do it in a way that We made it portable. We can run it on embedded systems with lower power, lower, resource capabilities. and that allowed us to put everything onto edge compute devices, so it didn't require a large workstation or systems in the cloud to process.

And we built purpose built hardware around that software that we could use in the field. So building ruggedized base stations that ran our computer vision software that also acted as a high precision base station for the drones. That was kind of one of the key products that we built at Skycatch that really made a difference in the customers because nobody else had that.

Everyone else, you have to. Fly the drone the same way, but then there's a dependency on getting that data to a [00:16:00] workstation or uploading gigs of data to the cloud for processing.

David SP-VG: So I think that's really insightful because I think a lot of people who are maybe starting up a company The hard problem that is core, you know, your budgets are limited or you might kind of do all the stuff around that and try to find some off the shelf solution, grab an API or some other existing kind of tool that does that.

But you guys identified early on that was the core thing and you had to own it and like really investing in that core thing. That's what really enabled you guys to get that portability you were describing and the ability to actually leverage it. Like if you had used an off the shelf solution, you don't think you would have been able to get the same outcomes.

David Chen: Yeah, I think it would have been a very different path. other companies have done it very successfully, uh, targeting different markets where there wasn't that need. But for us, because we had a focus on this tech, it opened up a lot of doors in, areas that really needed the portability and the quick access to data in the field.

David SP-VG: So the password is to Agtonomy because I just think it's so cool. when we talk about it, I'm [00:17:00] like, ah, it's so awesome. Just these, see these tractors driving around. And, tell me a little bit, maybe dig into that. Like now that you're at Agtonomy, what are some of the interesting challenges you guys have faced and, what has sort of been some of the more exciting things you've been able to crack into?

David Chen: Yeah, I think a common theme between Skycatch and Ektonomy is when the user interacts with these systems, they're using software that's super easy to use and super intuitive, which is making a super complex process like drones and high precision GPS and all of that stuff. You put it behind a good UX and that makes a huge difference.

Thanks. One of the key tech decisions at Ektonomy, and this was something that the team made right before I joined, was, using Flutter to develop all of their apps. I think, at the time, they also tried out, React Native, but ultimately went with Flutter, because it was just easier and quicker to get going.

And these guys are not You know, they're [00:18:00] not new to programming. They're all very smart, but focus on like more on the autonomy and embedded side. So seeing that, how they were able to quickly pick up Flutter and build a working app was something that was new to me. I had never used it before. Once I came on board and started building with it, I realized that.

It's incredibly powerful, and, I hired a team, some from previously at Skypatch, and some from other places that I knew, and they all have never worked with Flutter, brought them in, they were proficient in a matter of weeks, and that led us to building out now three apps on very different platforms, all sharing probably 70 percent common code base.

And letting us iterate and move really quickly while maintaining highly testable and reusable components.

David SP-VG: Yeah, you mentioned the portability thing before, right? in terms of even in the Skycatch story with your 3D mapping tech and making sure that you could get on these embedded systems andit's [00:19:00] interesting to hear you say about the 70 percent code reuse. I think when people think about cross platform or multi platform tools, they're thinking about like single code base, multiple apps, like same app, just different platform.

But is it correct to say that you guys are basically like actually distinct multiple apps, but where you're sharing a lot of core logic?

David Chen: Yep, exactly. So for us, we have a tablet app that's mainly for monitoring, controlling the fleet, planning your missions. and then we also have a web app, which is more for analytics and metrics. And we also have the actual on vehicle dashboard itself. This is when the operator is driving manually, they have a touch screen that they can see, telemetry and control the vehicle systems.

They're very, very different apps, but, because we use Flutter for all three, we have a shared SDK that's used across the board. We have shared widgets, we have shared data models. So all the foundational pieces. [00:20:00] is the same and a small group of engineers can maintain that. And then for the very specific platform things, we can build different things by composing these elements together and it just works.

David SP-VG: Yeah, it sounds like you guys took what I think is, the holy grail approach to these types of tools, which is not thinking about single code base, same app, multiple platforms, but what if we have an ecosystem of apps that are all architected, built consistently, and in a common, with a bunch of common shared libraries and set that up, so that your team can, your team then becomes really portable as well.

Like, are you guys able to just, easily hop across all three of those applications super easily?

David Chen: Yeah, totally. Nobody is assigned to working on a single app. they work across the apps and I found that to be really great. Really good and allows us to be very efficient. So someone might own the SDK stack across all the apps and if they need to make changes, that's [00:21:00] automatically available for all the other platforms.

David SP-VG: And the, if I'm correct, the one on the, tractor, that's not like an iOS iPad or something, is that, what operating system is that running?

David Chen: So for that one, it's an embedded Linux system. So that one also is Flutter and it just boots right into a Flutter application. When the system boots, it looks and feels exactly the same as our other apps. But you just never know what's underneath it.

David SP-VG: Yeah, that's, I never know what's underneath it, I feel like that's the new marketing campaign for Flutter right there. well done. no, I think, cause, one of the things I think we face when we look at some of these really transformative tools is you base your perception of it on prior mental models.

And you might look at a thing like React Native, or you might look at some of these other tools, and I think what you guys are doing that I've always been, like, that's really exciting about it is pushing it into these boundaries of no, this is truly, multi platform. This is, we're using multi platform.

Linux on a [00:22:00] ruggedized device on a tractor. We're not just talking about iOS and Android apps in your pocket. and would that have been a lot more difficult for you guys as a team, staffing, resourcing, hiring, if you had to hire, did it some other way?

David Chen: Absolutely. I think we would have needed a lot more engineers to actually build out everything that we have built out. And we would have to hire very specialized resources that are experts in that respective platform and language. Which meant,it would be much slower to build an app and we couldn't move people around to work on other parts of the system when there's needs.

David SP-VG: Yeah, that's great. And so it sounds like in both, both your experiences that you described, defining that 3d, core, core software and making a choice like Flutter, I'm sure they were like, Many, many, many other important technical decisions along the way. But it sounds like one of the things that you, you've prioritized in your roles is like making these like really important, high leverage strategic technology choices, right?

Choosing where to invest and what to [00:23:00] invest in to give your team The flexibility, portability, is a word we've said a couple of times, and sort of scalability to really be as productive as possible. how do you like, make those decisions? I mean, I guess to some extent the Flutter decision maybe predated you, but even still, as you consider down that path and making some of those choices, or when you're observing your teams making decisions around priority, or decisions around, like, critical technology choices, how do you as a team make those effectively?

David Chen: I think you have to take a step back and really look at what are the things that you will build or you will need, and really take that into consideration when you're architecting your apps. We're actually going through a pretty big refactor to bring a bunch of things up to speed. There were some tech debt that caused certain apps to lag behind, but being on Flutter, we were able to, build something new and port it over to the new apps or a different app with pretty minimal effort.

And [00:24:00] now they're all pretty much. Pretty much identical underneath. I think the Flutter ecosystem is moving really quickly and we're always finding new packages and new tools that will help us accelerate our development. So it's, it still feels. Relatively young in terms of Flutter itself. I know it's been around for almost 10 years now, but, it's still growing.

I see a lot of excitement around Flutter. I was actually just at a Flutter meetup, last night and met some of the Google team there that works on the Flutter frameworks and they're all super excited and just seeing. them talk about how the adoption of Flutter is really growing and accelerating all over the world is just really good signal that we've made the right choice.

David SP-VG: Well, I tend to agree with you for obvious reasons, and I think it's, uh, you know, the thing about Flutter is it's a high leverage tool, is like what I like to say about it. It's like all the things you were describing about the impacts [00:25:00] on your team. And there's a lot of things that you can do, and you've made choices elsewhere in your technology stack throughout.

All of your career experiences where it's like creating these leverage points for your team where you can kind of like unlock additional productivity. and Flutter, because you can, like, you know, anybody can work on any of these apps and the complexity in your organization, if each one of those apps is a different team staff, differently, different product manager, right?

Different engineers. And you have to coordinate them and talk through all the requirements. and really like at the end of the day, the user experience, like you said, you want to make the UI really, it's super critical you get that, but still like. There's a lot of work being done, I would think, on the autonomous side, like actually making the tractor actually be autonomous.

And in a way that's like the killer feature, but you need this front end and you don't want the front end development to get in the way of that other stuff. or be this extra strain and burden on the business.

David Chen: Yeah, absolutely. And I think for a lot of these industries, The UX is so important, not to make light of the autonomy side, that is an extremely hard problem and takes a lot of work to get [00:26:00] right, but what I found is no matter how good the underlying tech is, if you present the user that's using this as their livelihood and it, something breaks or doesn't work or costs them an extra hour to figure out how to do something, They're just going to throw it away and never use it again.

You have to make it easy to use. and that first impression of how well that app works is so important. these jobs, they, you know, every second counts, they can't spend time figuring out an app or if it's buggy or, you know, it's unintuitive. That's the most critical piece.

David SP-VG: Yeah, makes a lot of sense. these industries you're in, I find it fascinating because you don't really come across too many people that have built tech drones for mining, right? or even ag tech, I mean it's out there, it's a big, big sector, but You know, so many engineers gravitate towards, other types of things, whether it's e commerce or, online tools or whatever.

It's interesting to me that you've focused on these [00:27:00] connections and there's something about even those industries that there's some stigma to mining and even agriculture, you know, pesticides and things like, how did you get into those places? And how did you like, what's your perspective on getting into these types of industries that maybe an engineer out there might not be prioritizing when they're thinking about their career?

David Chen: Yeah. In terms of, at Skycatch, when I joined, I didn't know anything about construction or mining. mining, I think, to the outside perspective, it's like, oh, it's, it's dirty, it's hard work, there's toxic chemicals you hear about. You know, tailings dam collapses or failures that just pour toxic sludge down a valley and destroy the environment.

That's absolutely true. There's always these very negative aspects about these industries, but also everything depends on it. This is the core resources that Pretty much everything, the computers, the houses that we're in for ag, it's the food that [00:28:00] we eat every day. It's an extremely important. It's, it's something that will never go away.

And I think it's very important to have people working on improving these processes. How do we make. Mining safer. How do we make the environmental impact minimized by monitoring things more carefully and preventing these catastrophic failures? For ag, how do we minimize the use of chemicals in farming by using more intelligent systems that can do things mechanically instead of chemically?

David SP-VG: Very, very cool. There's definitely like a really interesting connection, right? Mining and agriculture, because there's like a supply chain that is so crucial to us living our everyday lives. You know, we're using Devices right now, or whether we're on the other end of this listening to it in our car, or on our airpods, that had materials that came out of those mines, 

and similarly, like, with food, or the wine we drink, [00:29:00] or whatever, that comes out of agriculture, and I just think it's really interesting, those threads of, like, building tech that helps to improve safety and quality. Uh, Efficiency,all those things to build a, build us a better world. I think that's super cool.

I'm, what, like, kind of parting thoughts or what, when you meet, an engineer out there or a team that's starting out and they're gonna build some complicated system, right? what advice do you have for those people that, they could follow to be more successful or optimize their chances of finding success with the tools they're building?

David Chen: I think the first thing is find what you truly love doing. And for me, what really unlocked that was drones. When I realized I had this passion for building software, hardware together into a system, and then once you find that, then I think it's to find how can you use your passion and skills to work on something that makes a big impact, a positive impact on the world. And [00:30:00] then if you can do these two things, and I, then I think whatever you're doing, you're going to find fulfillment and happiness every day doing it. I think that's really critical. I know a lot of people who work for large tech that, really hate their jobs, but they, maybe they get paid a lot of money to do it.

So they stick with it. But,I much rather know that what I'm working on is helping people on the other end who's using my product and hearing people visiting the customers, hearing them say how great this is and how much it's improved their lives. that's so important and, um, is all that I really need.

David SP-VG: Do you think that helps you write better code, or make better tech decisions? I'm curious, like, I mean, well, it gets to culture, right, as well. I mean, it's not only your personal passion, but all the things you're saying, like, are you happy, are you fulfilled, are you motivated, all those things, I think, maybe some of the reason those other people are not happy in a big tech job is maybe they're not aligned with, what they're doing, they're not excited about the product, but also sometimes, what Cultures can be not [00:31:00] great and it's not a fun environment or it's too much bureaucracy or whatever I'm, just curious if you think that drives better results or anecdotally if you've seen better outcomes on teams Where you have that enthusiasm and passion

David Chen: I think so. I think when you're passionate about the product you're working on and you care about the users of the product, then you get into that customer mindset of always thinking about how do you make this better for the end user? And I think that definitely drives better products because you're constantly thinking about, okay, when you understand what it takes to do their day to day work, you think about during their day when they're doing this, what is the pain point?

How can I make this process better and smoother? So maybe it doesn't help me make better code, but it helps me come up with better products. And ultimately, that's what I do. That's what's important for the users.

David SP-VG: Well, and it maybe drives you to go visit that mine. It's cold. It's uncomfortable. It's far away [00:32:00] But if you're passionate and really interested in solving these problems you're maybe as a team more likely to go out there and really experience it and seek to truly understand what the users are looking for.

We had a, we had Patricia Roller on here as an earlier podcast, and she talked a lot about the importance of customer insights, like how customer insights have to drive a lot of your decision making. And maybe if you don't care, you're not going to fight as hard to find those insights, right? exactly. 

well, this is great.

I don't know if you have any other parting thoughts. covered a lot of ground here. I think just incredible, experience and background. I think just an area and a space where, I think it's novel and super interesting. I mean the whole autonomous space, started with Waymo and, but getting into like where autonomy is going, I think people see, you hear about, you know, full self driving Tesla and you maybe hear about things like Waymo, but the idea that there's autonomous drones out mapping minds is extremely cool.

And thank you for sharing your stories today. And really appreciate you, helping us out.

David Chen: Yeah, absolutely. It's been a pleasure chatting with you.

David SP-VG: Where can people find out more about [00:33:00] Agtonomy? If you guys are hiring, or if, if somebody's out there happens to be looking for autonomous solutions for their tractors?

David Chen: Yep. We're hiring across the board on all the engineering teams. agtonomy. com.

David SP-VG: Alright, awesome. Well, thanks so much, David. Can't thank you enough.

David Chen: All right. Thank you. Take care.