Latest / Deep Tech Germany – AI, Robotics & Frontier Innovation by Startuprad.io™ / Twinsity: AI & Drones for Infrastructure Inspection

Transcript

Twinsity: AI & Drones for Infrastructure Inspection · Deep Tech Germany – AI, Robotics & Frontier Innovation by Startuprad.io™

• WEBVTT 00:00:04.877 --> 00:00:12.617 Welcome to StartupRad.io, your podcast and YouTube blog covering the German 00:00:12.617 --> 00:00:17.357 startup scene with news, interviews, and live events. 00:00:19.957 --> 00:00:23.897 Hello and welcome everybody. This is Joe from StartupRate.io, 00:00:24.057 --> 00:00:28.837 your startup podcast and YouTube blog from Germany, bringing you another episode 00:00:28.837 --> 00:00:33.237 in our cooperation with the European Enterprise Network and Hessen Trade & Invest. 00:00:33.437 --> 00:00:36.297 But before we get to that, I would like to welcome my guest, 00:00:36.497 --> 00:00:38.237 Fabian. Hey, how are you doing? 00:00:39.337 --> 00:00:45.697 Nice to meet you, Joe. And thanks really a lot for inviting me today to speak 00:00:45.697 --> 00:00:48.497 about our exciting business with you. 00:00:48.677 --> 00:00:51.737 Yes, we'll talk a lot about big things today. 00:00:52.257 --> 00:00:57.697 That is not a dirty joke. The very simple reason is you guys are inspecting 00:00:57.697 --> 00:00:59.297 large infrastructure pieces. 00:00:59.637 --> 00:01:03.217 Think bridges, think high voltage lines and stuff like that. 00:01:03.357 --> 00:01:06.897 But before we get into all of that and what you've done before, 00:01:07.097 --> 00:01:11.277 a little word from our sponsor. This recording is supported by Hessen Trade 00:01:11.277 --> 00:01:13.777 & Invest and the Enterprise Europe Network Hessen. 00:01:14.757 --> 00:01:19.877 This recording was made possible by HTII and the Enterprise Europe Network Hessen. 00:01:20.057 --> 00:01:24.657 These organizations have made tremendous contribution to helping startup businesses 00:01:24.657 --> 00:01:30.697 succeed and thrive, providing a range of services from helping to find grants 00:01:30.697 --> 00:01:31.897 to ongoing partnerships. 00:01:31.897 --> 00:01:37.717 By taking advantage of these resources, startup companies can network and develop 00:01:37.717 --> 00:01:41.137 innovative strategies for success on the international stage. 00:01:41.357 --> 00:01:44.217 The dedicated support of HTAI, 00:01:44.843 --> 00:01:49.443 And the Enterprise Europe Network Hessen is paramount in providing startup businesses 00:01:49.443 --> 00:01:51.663 with the tools for lasting success. 00:01:52.183 --> 00:01:57.483 Look for our dedicated sub-podcast in partnership with them called Tech Startups 00:01:57.483 --> 00:02:00.083 Germany or on our link tree. 00:02:01.083 --> 00:02:05.943 Now we got this out of the way. Fabian, very happy to have you here. 00:02:06.323 --> 00:02:11.363 And we may tell our audience you are a software engineer by training. 00:02:11.363 --> 00:02:16.743 And you walked me through a little bit your life, what you did before you founded 00:02:16.743 --> 00:02:18.643 your current startup, TwinCity. 00:02:20.243 --> 00:02:25.083 Can you try to do that again? I found it especially entertaining that you changed 00:02:25.083 --> 00:02:27.023 your employer without doing anything. 00:02:30.063 --> 00:02:35.063 Yes, right. So, yeah, as you said, originally I'm a software engineer. 00:02:35.103 --> 00:02:37.263 I studied computer science here in Kassel in Germany. 00:02:37.503 --> 00:02:40.423 It's between Frankfurt and Hannover. 00:02:42.323 --> 00:02:45.903 And started working at the company from 00:02:45.903 --> 00:02:49.143 my dad who founded a drone software company or 00:02:49.143 --> 00:02:52.363 drone hardware company as well called Ibotics and 00:02:52.363 --> 00:02:55.263 2010 I was working there as a software engineer 00:02:55.263 --> 00:02:59.903 focusing on let's say flight planning software for drones or for the drones 00:02:59.903 --> 00:03:07.343 that were built by Ibotics and yeah stated that industry now with founding Twinsity 00:03:07.343 --> 00:03:11.963 actually in the meantime I was I was at Fraunhofer Institute in Kassel working 00:03:11.963 --> 00:03:14.883 on renewable energy infrastructure, 00:03:15.343 --> 00:03:18.643 especially I was focusing on virtual power plants, which is like. 00:03:19.736 --> 00:03:22.736 Connecting decentralized renewable energy plants 00:03:22.736 --> 00:03:25.956 to provide stability for 00:03:25.956 --> 00:03:28.916 the grid but that's just let's say a different 00:03:28.916 --> 00:03:32.016 topic and yeah in 2019 00:03:32.016 --> 00:03:34.956 i started founding twinsity together with 00:03:34.956 --> 00:03:38.236 my father and still being the software 00:03:38.236 --> 00:03:41.296 developer for this part especially for the 3d development development 00:03:41.296 --> 00:03:45.216 of high resolution 3D models and the 00:03:45.216 --> 00:03:49.456 processing of data actually that is generated 00:03:49.456 --> 00:03:52.396 by drones so we are staying in the industry but concentrating more 00:03:52.396 --> 00:03:55.176 on the finding answers in the data that 00:03:55.176 --> 00:03:58.576 is captured by drones before we 00:03:58.576 --> 00:04:01.976 get into the finding answers part um 00:04:01.976 --> 00:04:05.976 you've been a software engineer at ibotics 00:04:05.976 --> 00:04:08.736 and that was the business of your dad and then 00:04:08.736 --> 00:04:11.896 it was brought by Leica Geosystems and 00:04:11.896 --> 00:04:15.096 basically that's how you changed your employer 00:04:15.096 --> 00:04:18.836 without ever leaving your current job right yeah 00:04:18.836 --> 00:04:21.676 the company ibotics as you mentioned was acquired. 00:04:21.676 --> 00:04:24.636 By Leica Geosystems which is a subsidiary of 00:04:24.636 --> 00:04:27.376 hexagon to quite a big player in the 00:04:27.376 --> 00:04:30.336 market for surveying equipment and now drones as 00:04:30.336 --> 00:04:34.396 well and laser scanners and data geospatial word 00:04:34.396 --> 00:04:38.556 data and so on and they bought 00:04:38.556 --> 00:04:45.856 us and so we let's say still was well a body was still like an own company within 00:04:45.856 --> 00:04:50.996 this complex or within this subsidiary Leica Geosystems so all developers were 00:04:50.996 --> 00:04:55.976 still at the company still developing further So yeah, the title changed, 00:04:56.276 --> 00:04:59.616 the company changed, but the job was the same. 00:05:02.440 --> 00:05:10.200 Basically, we may already tease that your dad is also your co-founder in TwinCity as well. 00:05:10.780 --> 00:05:15.500 Right. Yes. We founded the company together, kind of a family business in the beginning. 00:05:15.900 --> 00:05:21.200 Now it's not a family business anymore in the sense of that it's completely 00:05:21.200 --> 00:05:28.920 led by the family but it's now majoring into a, let's say, more diverse startup. up. 00:05:29.980 --> 00:05:34.940 There are not that many people who will actually watch this interview, 00:05:35.080 --> 00:05:39.020 but for those people who are watching this, we may tell them that you have pictures 00:05:39.020 --> 00:05:42.980 of high voltage lines and large windmills in the background. 00:05:43.100 --> 00:05:47.840 So that already sets the tone that some of the pieces of infrastructure we will 00:05:47.840 --> 00:05:49.180 be talking about pretty soon. 00:05:49.340 --> 00:05:52.200 Can you tell us, can you take us along the 00:05:52.200 --> 00:05:55.820 journey how you and your father actually 00:05:55.820 --> 00:05:58.660 founded Twin City and how did you come 00:05:58.660 --> 00:06:01.940 up with the name because um first thing everybody thinks 00:06:01.940 --> 00:06:06.100 smart cities huh yeah so 00:06:06.100 --> 00:06:09.240 we started founding the company after leaving 00:06:09.240 --> 00:06:12.140 the previous company Ibotics which was sold 00:06:12.140 --> 00:06:18.800 like several years after it was sold and in the time during the previous company 00:06:18.800 --> 00:06:23.440 we recognized quite heavily the problem or the the challenge of companies using 00:06:23.440 --> 00:06:28.280 drones that they don't really know what to do with the images being collected 00:06:28.280 --> 00:06:30.200 by the drone. So what comes next? 00:06:30.460 --> 00:06:34.380 Drones are just flying cameras producing thousands of images for one bridge, 00:06:34.500 --> 00:06:40.100 for example, but finding answers in these images, that's the challenge, 00:06:40.180 --> 00:06:42.480 the bottleneck, and that's still the bottleneck. 00:06:42.500 --> 00:06:46.340 So that was the reason why we decided to. 00:06:47.697 --> 00:06:52.477 Let's say investigate this problem and try to find a solution to that, 00:06:52.657 --> 00:07:01.677 that will analyze those huge amounts of images and generate high resolution 3D data out of that. 00:07:01.937 --> 00:07:06.597 So producing like a Google Earth version of the asset but in much higher quality, 00:07:06.857 --> 00:07:08.837 that was the overall game at the beginning. 00:07:09.037 --> 00:07:15.017 So producing a digital twin, so to say, of an asset that is inspected by the drone. 00:07:15.197 --> 00:07:20.277 That was the the major component that we developed as the first, 00:07:20.337 --> 00:07:22.597 let's say, starting point for Twinsity. 00:07:22.757 --> 00:07:26.097 So how we can create a digital twin out of the images. 00:07:26.857 --> 00:07:34.777 And when you're talking about an asset coming from the asset management background, 00:07:35.137 --> 00:07:41.317 not physical assets, but talking about bonds and funds and shares here, 00:07:41.517 --> 00:07:43.337 that is something different. 00:07:43.337 --> 00:07:46.617 We are always talking about structures when you talk about asset, 00:07:46.817 --> 00:07:50.577 plus it always implies that somebody owns it. 00:07:50.677 --> 00:07:56.957 And we can rightly assume the person or the entity who owns this asset is actually 00:07:56.957 --> 00:07:59.757 the one who is hiring you. 00:07:59.897 --> 00:08:06.277 And with you, it's not flying the drone around, but basically they drop you a lot of data. 00:08:06.877 --> 00:08:10.797 Yes, right. Right. So when we talk about ACID, as you said, it's like about 00:08:10.797 --> 00:08:15.997 big infrastructure objects like a bridge, high voltage power lines, 00:08:16.177 --> 00:08:19.557 industrial facilities, oil and gas platforms, but also buildings. 00:08:19.597 --> 00:08:22.897 So everything that needs to be inspected on a periodic basis, 00:08:23.077 --> 00:08:27.157 this can be, so to say, used with our platform. 00:08:27.157 --> 00:08:32.417 And you're right, our major customers are the asset owners and operators, 00:08:32.737 --> 00:08:37.177 these that are responsible for ensuring the safety, 00:08:37.457 --> 00:08:42.157 reliability of the assets that are interested in keeping them, 00:08:42.157 --> 00:08:48.917 let's say, alive and no need to turn them off, like for offshore platforms, 00:08:49.157 --> 00:08:54.077 for example, or for road closures or even more catastrophic events. 00:08:54.077 --> 00:08:57.737 So these are the customers that we have. 00:08:57.997 --> 00:09:02.197 Those customers, like asset owners, already have drone programs in place. 00:09:02.377 --> 00:09:06.697 So they already utilize drones for capturing the data of their assets. 00:09:07.677 --> 00:09:12.837 And so they are using our software and providing us with the images they already collect. 00:09:13.137 --> 00:09:16.457 And so we are not a service provider. We don't own the drones. 00:09:16.617 --> 00:09:20.077 We don't fly the drones. but we deliver the operating system, 00:09:20.177 --> 00:09:24.757 the software side, for those asset owners that wants to get insights from the 00:09:24.757 --> 00:09:25.817 data they already collect. 00:09:27.064 --> 00:09:30.904 And just to be sure, what do you actually get? 00:09:30.984 --> 00:09:37.684 Do you already get 3D images or do you get just simple digital photos and videos 00:09:37.684 --> 00:09:39.524 and you have to do all the stuff with it? 00:09:40.084 --> 00:09:46.184 Usually we get the images, like the data that is captured by the drones directly. 00:09:46.444 --> 00:09:52.264 So the raw data that the drones capture. So this is especially RGB images, 00:09:52.544 --> 00:09:58.904 but also sometimes thermal data, like thermal images, but usually RGB for the 00:09:58.904 --> 00:10:00.024 visual inspection side. 00:10:00.744 --> 00:10:06.084 And we then can generate those 3D models on our own internally. 00:10:06.084 --> 00:10:12.084 So we only need images and we create the 3D model and we analyze the images 00:10:12.084 --> 00:10:14.264 for defects, basically. 00:10:15.904 --> 00:10:23.384 And as we've been speaking before, what I had in mind and many of our audience 00:10:23.384 --> 00:10:27.444 will have in mind, well, it's a bridge, it stands there and that's it. 00:10:27.444 --> 00:10:32.444 But actually you highlighted that there are quite some challenges for infrastructure 00:10:32.444 --> 00:10:36.924 and not like the Romans who built bridges who are still up and running today. 00:10:37.204 --> 00:10:42.144 The bridges we are, or we have been constructing in the past, 00:10:42.304 --> 00:10:46.024 have a planned lifespan of around 60 years, right? 00:10:46.704 --> 00:10:50.384 Right, yeah. And they are built probably in the 60s and 70s, 00:10:50.384 --> 00:10:51.664 the majority of all bridges. 00:10:52.104 --> 00:10:56.984 So now they are more and more reaching the end of their design life, 00:10:56.984 --> 00:11:00.344 time which means that we 00:11:00.344 --> 00:11:08.044 now need to deploy more sophisticated and ensuring safety measures and repair 00:11:08.044 --> 00:11:14.644 works and maintenance activities to prolong the lifetime of the bridges compared 00:11:14.644 --> 00:11:18.824 to the design time so it's really important that we. 00:11:20.286 --> 00:11:25.986 Say, strengthen them in order to reuse them or to still be able to use them 00:11:25.986 --> 00:11:28.466 over the next couple of years and even longer, 00:11:28.586 --> 00:11:33.766 rather than, let's say, creating or building new bridges, which is much more 00:11:33.766 --> 00:11:38.146 heavy when it comes to the natural resources, human resources, and so on and so forth. 00:11:39.606 --> 00:11:47.566 Every commuter here in the Rhein-Main area can quite talk a lot about bridges 00:11:47.566 --> 00:11:51.746 being not maintained as much as they should be. 00:11:51.866 --> 00:11:55.406 What comes to mind is Schiersteiner Brücke, Salzbachtalbrücke, 00:11:55.526 --> 00:12:01.966 two really big obstacles where people needed to drive around for like 30 minutes 00:12:01.966 --> 00:12:04.266 additional one way or stuff like that. 00:12:04.346 --> 00:12:09.486 That wasn't fun and with all the rates here in the Rhein-Main area for quite 00:12:09.486 --> 00:12:15.786 some time because first one bridge was closed and there was one rebuilt just next to it. 00:12:15.866 --> 00:12:20.726 And then another one on the same way to Frankfurt was closed for some time. 00:12:20.886 --> 00:12:25.706 They actually had an inspection truck and the inspection truck data was so severe, 00:12:25.846 --> 00:12:30.466 they had to lock down the bridge and then had to take a lift and lift this truck 00:12:30.466 --> 00:12:34.666 off the bridge because it was not safe enough for the truck to continue. 00:12:34.846 --> 00:12:38.966 So there are some challenges out there. When we talked before, 00:12:39.126 --> 00:12:44.426 you said there is a life expectancy of those infrastructures. 00:12:44.586 --> 00:12:49.166 How long, for example, would something like a high voltage line last? 00:12:49.306 --> 00:12:54.866 Or as you said, like an oil drilling rig or something like this. 00:12:54.946 --> 00:12:57.626 Do you have any idea about those assets? 00:12:59.326 --> 00:13:04.726 That's really depending on the type of assets. because like a power voltage 00:13:04.726 --> 00:13:12.166 tower or like a utility tower can be made of steel, of concrete or of wood. 00:13:12.326 --> 00:13:18.706 So it's really depending on the material being used as well as the weather conditions at the site. 00:13:18.826 --> 00:13:25.666 So, for example, in California, there will be different weather situations compared 00:13:25.666 --> 00:13:30.886 to what we have in Europe, for example, in Germany and Scandinavia. 00:13:30.886 --> 00:13:37.486 So we can't really talk about the expected lifetime of those assets, 00:13:39.046 --> 00:13:43.286 let's say, defining a clear number of years. 00:13:43.506 --> 00:13:48.586 But for these kind of assets, it's much more important to still know exactly 00:13:48.586 --> 00:13:54.226 about the condition of the assets because of these weather situations that are quite severe. 00:13:54.766 --> 00:13:59.086 And so the assets like the power poles will. 00:14:01.850 --> 00:14:07.650 Deteriorate and will be more affected to those weather situations which means 00:14:07.650 --> 00:14:15.290 like like um there can be a lot of damages and potential defects that will be quite, 00:14:15.990 --> 00:14:18.830 critical to those assets in order to maybe 00:14:18.830 --> 00:14:22.110 it needs to be or which means that those 00:14:22.110 --> 00:14:25.030 assets may might need to be replaced or 00:14:25.030 --> 00:14:28.390 even can um collapse which 00:14:28.390 --> 00:14:31.470 will turn off the overall and utilities so 00:14:31.470 --> 00:14:35.230 to say in that region so it's really important that we 00:14:35.230 --> 00:14:38.150 know about the condition up front in order to 00:14:38.150 --> 00:14:41.270 maintain those power poles for example or 00:14:41.270 --> 00:14:45.090 even like oil and gas refineries offshore platforms 00:14:45.090 --> 00:14:51.930 so it's not just about the time let's say how long or what's the expected lifetime 00:14:51.930 --> 00:14:58.110 from the beginning on from the design time but it's just really important in 00:14:58.110 --> 00:15:01.330 now over the next years to really know exactly about the condition. 00:15:03.870 --> 00:15:10.770 I see. Before we get into a little bit more what you guys are doing and about the drones, 00:15:11.530 --> 00:15:16.950 what I found fascinating is he told me how those high voltage lines are actually 00:15:16.950 --> 00:15:19.110 have been inspected in the past. 00:15:19.230 --> 00:15:22.770 That means somebody rented a helicopter and 00:15:22.770 --> 00:15:27.850 you put like half a dozen engineers in this helicopter and they flew along those 00:15:27.850 --> 00:15:33.450 high voltage power lines which I assume was quite pricey even though what I 00:15:33.450 --> 00:15:41.130 had in mind was something like an open an open door in a helicopter somebody tied to the, 00:15:41.750 --> 00:15:46.570 helicopter with all this gear and then looking out with the binoculars onto 00:15:46.570 --> 00:15:48.330 the different high voltage poles. 00:15:48.550 --> 00:15:53.570 Was there something close to realistic that that happened in the past? 00:15:54.590 --> 00:15:57.410 I mean, that's exactly how it works right now. 00:15:57.570 --> 00:16:04.150 And not just in the past, that's still how many companies do the inspection activities. 00:16:06.954 --> 00:16:13.194 Because it's, I mean, that's still the safest way to do it compared to needing 00:16:13.194 --> 00:16:17.814 scaffolding or some climbers that will climb on the tower. 00:16:18.014 --> 00:16:21.394 It's still better to do it, let's say, not directly at the tower, 00:16:21.494 --> 00:16:22.774 but with a bit of distance. 00:16:23.034 --> 00:16:27.774 But it's quite expensive to hire or to rent helicopters to do the job. 00:16:27.874 --> 00:16:32.714 It's quite expensive to have six or, like you said, different engineers sitting in the helicopter. 00:16:32.714 --> 00:16:35.934 And I mean with an helicopter you 00:16:35.934 --> 00:16:38.694 can't do that many different power poles per 00:16:38.694 --> 00:16:45.614 day because you really need to be far and near to the power pole and still you 00:16:45.614 --> 00:16:52.534 can't get to the maximum detail so to say that you probably would need to do 00:16:52.534 --> 00:16:57.854 so that's where the drones get in to automate this process to capture the data. 00:16:58.474 --> 00:17:01.834 Before we get into that, just one question to confirm. 00:17:02.094 --> 00:17:07.774 Are there people whose job it is to climb on high voltage poles? 00:17:08.534 --> 00:17:12.614 Again, this is not the favorable way. 00:17:13.874 --> 00:17:19.774 Yeah, but for repairs or something like that. Yeah, that's definitely happening. 00:17:20.134 --> 00:17:24.534 Yeah. I'm curious if those guys get life insurance anywhere. 00:17:25.234 --> 00:17:29.574 Anyway, sorry, you got a little bit off the topic here so that means people 00:17:29.574 --> 00:17:33.274 are flying drones they they're likely a little bit bigger than your average 00:17:33.274 --> 00:17:39.494 drone you have at home and they capture a lot of pictures which means it's cheaper. 00:17:40.514 --> 00:17:46.574 Does it also mean your clients provide you more frequently with updates with 00:17:46.574 --> 00:17:55.514 pictures with video with infrared recordings to put in your database to maintain 00:17:55.514 --> 00:17:59.294 the digital twin in your database of the asset? 00:17:59.634 --> 00:18:04.814 Absolutely. That's the beauty about the drones. They capture the data autonomously. 00:18:04.894 --> 00:18:07.554 That means you don't need to pilot anymore. 00:18:08.154 --> 00:18:11.294 I mean, that's changing quite a lot now that 00:18:11.294 --> 00:18:15.054 drones will really capture the data automatically with 00:18:15.054 --> 00:18:18.094 a predefined route without any pilot being 00:18:18.094 --> 00:18:20.754 involved anymore so we always get the 00:18:20.754 --> 00:18:23.854 same number of pictures for the same route and so 00:18:23.854 --> 00:18:26.714 this this route can be done every day 00:18:26.714 --> 00:18:31.074 every week every month every three months depending on what the customer wants 00:18:31.074 --> 00:18:37.674 or what the asset owner is yeah relying on what they what they want to do so 00:18:37.674 --> 00:18:45.234 for us it doesn't matter if it's a weekly base a daily base a monthly base but still In every case, 00:18:45.254 --> 00:18:51.774 it's much higher frequency compared to hiring a helicopter that does it every two years once, 00:18:51.934 --> 00:18:54.054 so to say, or one time every two years. 00:18:54.334 --> 00:18:59.354 With drones, you get really much more frequent inspections. 00:19:01.922 --> 00:19:05.102 Mm-hmm. I see. And now comes the magic part. 00:19:05.262 --> 00:19:09.962 What are you guys actually doing with all those pictures, videos, 00:19:10.342 --> 00:19:12.442 infrared recordings, and so on and so forth? 00:19:13.302 --> 00:19:21.162 Personally, I would assume you have very capable AIs who then extract the necessary data from it. 00:19:21.242 --> 00:19:25.802 You have a big database where you put all of this in, and then you do have a 00:19:25.802 --> 00:19:32.162 digital twin in your database where you would, I think, share again the data 00:19:32.162 --> 00:19:35.202 with the asset owner, with the client? 00:19:35.702 --> 00:19:40.542 Well, that was a really good summary of our product or of our software. 00:19:40.722 --> 00:19:45.502 So basically, there are two different components involved in the process. 00:19:46.002 --> 00:19:50.382 That is, as you said, it's completely correct. It's a digital twin. 00:19:50.542 --> 00:19:54.002 That's one of the proprietary technologies that we developed over the years 00:19:54.002 --> 00:20:00.142 to take all the data from the drones from the asset owners and produce a high 00:20:00.142 --> 00:20:05.482 resolution 3D model of the object to be inspected like of the overall air bridge 00:20:05.482 --> 00:20:10.322 or of 10,000 of power poles or of the overall offshore oil platform. 00:20:12.158 --> 00:20:16.978 We just take the data and create this high-resolution 3D digital twin out of that. 00:20:17.478 --> 00:20:22.238 And this digital twin can be visualized on any device with our web platform. 00:20:22.418 --> 00:20:24.098 So you can use it on a tablet. 00:20:24.578 --> 00:20:28.318 The engineer can use it on a smartphone, on a laptop, in the office. 00:20:28.418 --> 00:20:35.078 No matter where you are, it's streamed over the web into your web browser, so to say. 00:20:35.238 --> 00:20:40.658 So our software is a cloud-based software solution. The second part, 00:20:40.818 --> 00:20:47.958 as you mentioned, is the AI to autonomously detect defects and issues up front in the data. 00:20:48.158 --> 00:20:54.058 So what this means is we take all the images and our AI runs over that to detect, 00:20:54.298 --> 00:20:57.798 for example, on bridges, cracks or rust. 00:20:58.018 --> 00:21:02.838 That's just two examples of many different defect types. So the software runs 00:21:02.838 --> 00:21:07.678 over all images, detects cracks, and so the engineers of the asset owners will 00:21:07.678 --> 00:21:10.238 eventually get a list of all cracks we found. 00:21:10.438 --> 00:21:16.518 And so they can review crack by crack and give us feedback about if this really 00:21:16.518 --> 00:21:18.018 is a crack or if it's not a crack. 00:21:18.138 --> 00:21:22.518 So with every inspection we do or our customers do within the platform, 00:21:22.798 --> 00:21:24.738 our software gets better and 00:21:24.738 --> 00:21:29.118 better because we always ingest the data back into the training of the AI. 00:21:29.938 --> 00:21:37.418 Talking about cracks and bridges here, how big or how small can they be in order 00:21:37.418 --> 00:21:40.838 to be detected by your software and admittedly, of course, 00:21:40.898 --> 00:21:45.138 a lot is dependent on the quality of pictures you get delivered from the drones. 00:21:46.058 --> 00:21:51.998 So actually, there's no limitation in the size. So the bigger the better for 00:21:51.998 --> 00:21:53.398 our software to detect, obviously. 00:21:53.698 --> 00:21:55.578 That will be a takeout for Christmas. 00:21:59.118 --> 00:22:02.098 I figured the bed and the cracks Fabian, there you go. 00:22:03.478 --> 00:22:10.798 Yeah, but I mean, let's say in general, the cracks can be up to one millimeter 00:22:10.798 --> 00:22:13.298 in size. So that's maybe the lower limit. 00:22:14.634 --> 00:22:18.494 Which are the more relevant cracks to detect actually for the 00:22:18.494 --> 00:22:21.614 software because bigger cracks will be detected even without 00:22:21.614 --> 00:22:24.334 the software you can already see it when you 00:22:24.334 --> 00:22:27.254 have a look at the bridge being on side so it's important 00:22:27.254 --> 00:22:30.614 that we especially support those small 00:22:30.614 --> 00:22:33.674 cracks one millimeter in 00:22:33.674 --> 00:22:37.754 width for example that the human 00:22:37.754 --> 00:22:40.794 egg i can't see when you be on side so that's 00:22:40.794 --> 00:22:43.454 really crucial that we recognize or that we did 00:22:43.454 --> 00:22:46.654 detect those issues and we can so up 00:22:46.654 --> 00:22:49.874 to one millimeter as i said we do already have 00:22:49.874 --> 00:22:53.474 a pretty good understanding of what 00:22:53.474 --> 00:22:56.474 you guys doing where you guys going i 00:22:56.474 --> 00:23:00.154 do have a few questions for you on what 00:23:00.154 --> 00:23:03.394 you'll be doing in the future um we've been 00:23:03.394 --> 00:23:06.314 talking about such a lot of about structures and 00:23:06.314 --> 00:23:12.614 assets is there one you would like to inspect with the software what comes to 00:23:12.614 --> 00:23:18.174 mind are really iconic items like the statue of liberty the hoover dam the eiffel 00:23:18.174 --> 00:23:23.394 tower the brangenberg gate so the eiffel tower would be really really interesting because, 00:23:24.034 --> 00:23:33.194 it's quite a big steel structure that is not really easy to model as a or to to process as a 3D model. 00:23:34.014 --> 00:23:37.354 Because it's that, I mean, it's like a really big power pole, 00:23:37.474 --> 00:23:39.814 actually, or like a huge transmission tower. 00:23:40.034 --> 00:23:46.634 But that size of an object is would be really challenging. 00:23:46.814 --> 00:23:51.534 And thus, it would be really interesting for me to see how the software is capable 00:23:51.534 --> 00:23:53.674 of doing exactly this kind of structure. 00:23:54.074 --> 00:23:56.294 So the Eiffel Tower it is. 00:23:58.414 --> 00:24:05.254 We do have a lot of our audience in Europe but also something like, 00:24:05.254 --> 00:24:08.834 I don't know, 50,000 listeners a month in the US and abroad. 00:24:11.334 --> 00:24:21.814 Where are you guys available and is there like a real limit to where the structure, 00:24:22.134 --> 00:24:23.314 the asset could could be. 00:24:23.494 --> 00:24:25.914 So right now we are quite, 00:24:28.254 --> 00:24:33.014 focusing on Europe, so especially on Scandinavia, on Netherlands, 00:24:33.414 --> 00:24:40.514 on Germany and the overall DACH region, but also now heading a bit into southern Europe. 00:24:41.194 --> 00:24:50.654 And over the year, we will definitely, we are looking into scaling into the 00:24:50.654 --> 00:24:54.854 US or North America and also Latin America as well. 00:24:54.854 --> 00:24:59.574 So we are about to open an office over there in the next couple of months, 00:24:59.734 --> 00:25:01.954 hiring some people over there. 00:25:02.574 --> 00:25:08.574 And yeah, in the end, it doesn't matter where the acid is. They are quite similar defects. 00:25:09.414 --> 00:25:13.574 A bridge in US looks quite the same like a bridge in Europe. 00:25:13.694 --> 00:25:18.374 So there will be also cracks, rust, spalling, different defect types on bridges, 00:25:18.554 --> 00:25:23.074 as well as the power poles are also made of wood, steel or concrete. 00:25:23.074 --> 00:25:25.154 So that's quite comparable to Europe. 00:25:25.354 --> 00:25:29.594 So for all the from a software perspective, it doesn't matter where the asset 00:25:29.594 --> 00:25:34.914 is I see and when you talked about a Scandinavia what came to mind is the very 00:25:34.914 --> 00:25:40.214 severe weather there There could be some places where it gets minus 40 degrees Celsius, 00:25:40.814 --> 00:25:45.854 No idea what this is in Fahrenheit, but it's freaking freaking cold and you 00:25:45.854 --> 00:25:49.694 have high voltage lines and stuff like this They're bridges Plus, 00:25:50.034 --> 00:25:55.254 when you talk about the Netherlands, what came into mind are all the dam structures there. 00:25:55.394 --> 00:25:59.434 For everybody who's not familiar with the country, it's called the Netherlands 00:25:59.434 --> 00:26:05.874 because a large share of this country is reclaimed from the ocean and the actual 00:26:05.874 --> 00:26:09.254 level is already under the normal sea level. 00:26:09.374 --> 00:26:15.394 So they have to work with a lot of dams, a lot of dewatering facilities in order 00:26:15.394 --> 00:26:19.374 to keep the country alive. lives. Is that something you're working on? 00:26:20.494 --> 00:26:29.694 For sure. So for bridges in Scandinavia, we are quite heavily working with engineering 00:26:29.694 --> 00:26:35.654 companies over there that are giving us the expertise on specific damage types 00:26:35.654 --> 00:26:38.414 that are really known for Scandinavia, 00:26:38.494 --> 00:26:41.814 but they're not happening in the rest of the world, for example, 00:26:41.954 --> 00:26:44.214 or that are really specific for those regions. 00:26:44.294 --> 00:26:51.014 So So that's exactly what we have in mind with our platform to be agnostic in 00:26:51.014 --> 00:26:53.274 the sense that companies in 00:26:53.274 --> 00:26:58.594 Scandinavia can retrain our platform for their specific defects without us, 00:26:59.547 --> 00:27:05.827 being involved into that process. So their experts know exactly about the problems 00:27:05.827 --> 00:27:08.467 they have in Scandinavia, but we don't. 00:27:08.507 --> 00:27:10.887 But they know exactly, so they 00:27:10.887 --> 00:27:16.207 can retrain our platform to be capable of detecting those issues as well. 00:27:16.467 --> 00:27:23.207 So Scandinavian version of the platform, so to say, Scandinavian Bridge version of our software. 00:27:23.327 --> 00:27:28.487 That's what we have in mind to develop, so to be that open and adaptive platform 00:27:28.487 --> 00:27:32.887 where our customers can specify what they want to detect, 00:27:33.427 --> 00:27:38.467 and the platform adaptively gets better for their specific problems. 00:27:38.827 --> 00:27:45.467 I see and before we get into that minus 40 degrees fahrenheit are minus 40 degrees 00:27:45.467 --> 00:27:47.567 celsius so everywhere it's freaking cold. 00:27:48.867 --> 00:27:54.587 That said I do believe if you have your AI trained for Scandinavia with all 00:27:54.587 --> 00:28:00.427 the severe the weather out there, specific defects on bridges, and so on and so forth. 00:28:00.547 --> 00:28:03.747 Maybe other countries like Canada would also profit from that. 00:28:04.307 --> 00:28:07.647 Talking about all those options for international expansions, 00:28:07.747 --> 00:28:10.307 are you guys actually open to talk to the investors? 00:28:10.967 --> 00:28:19.487 Yes, we are actively raising our funding right now, which shall be done until end of Q2 this year. 00:28:20.527 --> 00:28:26.167 I see. and you guys are also looking for talented people to join you? Absolutely. 00:28:26.867 --> 00:28:30.087 We are looking into expanding our team right now. 00:28:30.187 --> 00:28:36.247 As I said, one part is US office and US expansion, but on the same side, 00:28:36.327 --> 00:28:41.187 we are looking for really great software engineers in the cloud architecture 00:28:41.187 --> 00:28:48.047 and computer vision departments as well as great business development managers 00:28:48.047 --> 00:28:51.987 and sales people people for scaling our business in Europe. 00:28:53.056 --> 00:28:56.476 Are there mostly remote or all remote positions included? 00:28:57.036 --> 00:29:02.236 We are a remote-first company, so we are looking especially for remote jobs. 00:29:02.836 --> 00:29:10.936 We have an office facility here in Kassel, but our team is quite spread around 00:29:10.936 --> 00:29:13.216 Europe and the U.S. already. 00:29:14.616 --> 00:29:20.256 I see. For everybody who would like to learn more, you can go down here in the show notes. 00:29:20.256 --> 00:29:24.076 They will be linked to Fabian's LinkedIn profile as well as the company website 00:29:24.076 --> 00:29:29.816 as well as your careers website and since this is in cooperation with Hessen Trade & Invest and the, 00:29:30.456 --> 00:29:36.416 Enterprise Hero Network here, there is your option to address the decision makers 00:29:36.416 --> 00:29:37.996 here in the state of Hessen, 00:29:38.856 --> 00:29:40.356 ideas, concerns. 00:29:42.076 --> 00:29:43.496 Things they could improve. 00:29:43.736 --> 00:29:46.876 Do you have anything to address here or questions to ask? 00:29:46.876 --> 00:29:52.016 So, as you know, we are based in Kassel, which is quite nowhere of Hessen, 00:29:52.216 --> 00:29:57.916 compared to Frankfurt, which is the central hub on Darmstadt and Wiesbaden and 00:29:57.916 --> 00:30:04.056 probably other cities that are more known to be startup friendly. 00:30:04.656 --> 00:30:10.596 I mean, Kassel is trying to do things for startups, but I think there is a lot 00:30:10.596 --> 00:30:18.096 of potential to optimize the landscape for startups. Just one example is the it's quite. 00:30:19.487 --> 00:30:24.767 I wouldn't say impossible, but it's really challenging to get an office here in Kassel for a startup. 00:30:25.007 --> 00:30:31.147 You have some like a campus or you have like a building for some startups, 00:30:31.207 --> 00:30:36.087 but this is not really, let's say, open to everybody. 00:30:36.587 --> 00:30:38.127 It's really close to the university. 00:30:38.747 --> 00:30:45.307 And so I think the ecosystem for university spinoffs is quite good in Kassel 00:30:45.307 --> 00:30:47.687 already or in Hessen, especially in Kassel. 00:30:47.687 --> 00:30:53.767 But for other startups, I think the ecosystem can be more, 00:30:53.987 --> 00:31:01.447 yeah, I mean, it can be broader, it can be more supportive for the startups 00:31:01.447 --> 00:31:06.987 with maybe mentorships, but also something, 00:31:07.167 --> 00:31:12.807 yeah, quite easy like an office building where you can now finally start building your company. 00:31:12.807 --> 00:31:18.847 I think that's what still is, let's say, to be optimized. 00:31:20.887 --> 00:31:26.407 I see. So, Fabian, it was a pleasure talking to you. Thank you very much. 00:31:26.547 --> 00:31:28.567 Good luck for your expansion plans. 00:31:28.727 --> 00:31:33.207 And when you successfully launched in the US, we will report about it again. 00:31:33.827 --> 00:31:38.607 Thank you very much for inviting me again. It was a pleasure talking to you. 00:31:38.607 --> 00:31:44.887 I think yeah I liked it a lot and appreciate that so let's speak again once 00:31:44.887 --> 00:31:49.887 we are in us sure my pleasure have a good day bye bye bye bye. 00:31:54.160 --> 00:32:20.445 Music.