Artificial Intelligence made in Bremen, Germany (Dr. Sirko Straube, DFKI and Roland Becker, JUST ADD AI)

00:00:00: Music.

00:00:06: Boris Felgendreher: Hello and welcome to "Go Global! Bremen Business Talks", a podcast about the businesses, the technologies and the people that connect the cities of Bremen and Bremerhaven to the rest of the world. "Go Global! Bremen Business Talks" is a joint project by BremenInvest,

00:00:20: the Bremen chamber of commerce and the senator for Economic affairs, labour and Europe.

00:00:24: I'm your host, Boris Felgendreher, and today's episode is about artificial intelligence made in Bremen. As it turns out, Bremen is a key centre for AI research in AI business.

00:00:36: With a globally recognised ecosystem of the academia, research institutions, companies and talent specialised in artificial intelligence.

00:00:45: My guests today are Dr. Sirko Straube, a leading researcher at DFKI, the German research centre for artificial intelligence in Bremen

00:00:52: and Roland Becker, founder and CEO of JUST ADD AI, an exciting AI startup from Bremen.

00:00:58: Roland is also the founder of BremenAI, a network that brings local AI companies and institutions together. Let's jump right in!

00:01:07: Boris Felgendreher: Roland, Sirko, welcome to "Go Global! Bremen Business Talks" podcast, thanks for being on the show guys. Both: Thanks for having us, in our own studio, awesome. Boris Felgendreher: Yes, for me it's a premier

00:01:17: on so many levels. For number one, you know, since we started this podcast a few months ago we've done 100-percent of the episodes remotely via zoom and this is the first time we're doing it 

00:01:27: live again, it feels very refreshing, this is sort of good already. Roland Becker: It has a natural touch to it. We could even touch each other. Boris Felgendreher: But there is enough distance between us.

00:01:39: That's number one, number two:

00:01:42: Video, right. First time we're doing this on video, fantastic. You use this all the time right? Dr. Sirko Straube: We do it all the time, yes. You will forget about it.

00:01:55: So I mean our goal is 

00:01:58: that everyone tries to be as natural as possible and that we just talk about artificial intelligence and other stuff.

00:02:06: Roland Becker: I think many of the people who listen to "THINK REACTOR" are actually listening to it just a podcast. I think some of them might not be aware that we do have a video cast also available on YouTube

00:02:20: so yeah, just do your natural thing and forget about the cameras.

00:02:25: Boris Felgendreher: Yeah, and of course the third level is to be invited in a guest studio, which never happens. I also do a podcast for the German Logistics Association

00:02:34: I've done over 100 interviews, not a single time has somebody invited me and has a full fletched 

00:02:38: audio and video studio. Professional stuff here. Very nice. Well, let's get started for the audience of the "Go Global! Bremen Business Talks" podcast. Let's start with you, Roland, what's your background, I learned that you are from

00:02:53: Berlin.

00:02:54: And now you're in Bremen, that's already a good decision to be made. Roland Becker: I grew up in Berlin, in the western part of Berlin

00:03:05: and I enjoyed my time in Berlin very much, I still love Berlin, when I come to Berlin I still feel at home.

00:03:14: I still smell the air and feel the vibes. But I left Berlin just after finishing high school

00:03:23: And I've been closer to the coast because I'm a passionate surfer.

00:03:29: So I went to Kiel initially and then for a couple of years I was in Kiel and then in Hamburg, then in Cape Town.

00:03:37: And then back in Hamburg, which turned out to be a very nice of fusion of being close to the coast,

00:03:43: also being a larger City than Kiel and in 2012 I changed from Hamburg to Bremen.

00:03:54: At first, it was for family reasons, because my wife is from the region and we had two kids coming

00:04:07: and support was needed from the wider family and she is very close with her family, so we decided to move to Bremen, so I could continue to work 

00:04:19: And then after working remotely for a couple of years in Hamburg or Düsseldorf and other places, I started to look for opportunities in Bremen.

00:04:29: And came to the conclusion that it could be a very cool idea to build an AI company here in Bremen. Boris Felgendreher: What interested you in AI?

00:04:42: Roland Becker: Yeah, so I talked about my personal things, so originally I started studying political science and economics and also business administration

00:04:54: but I did that during the time of the new economy

00:05:00: So in the late 90s and one of my best friends was a programmer at a developer

00:05:07: and so actually I think in '98 or '99 we had our first company and so we were still studying of course.

00:05:17: So we had our first company, did our first successful projects, you know, delivered on them and they went into production,

00:05:24: and at that time, it was like, if you were able to deliver some sort of digital project.

00:05:30: You immediately had 5 job offers on the table. Boris Felgendreher: It was that time.

00:05:38: Roland Becker: Yeah so actually I decided at that time I would

00:05:44: join a larger company and not one of the cool hot startups, which turned out to be a good decision because one year later, in 2000

00:05:53: the economy crashed and all the little companies were a little bit

00:05:58: in trouble, some of them still exist though, some of the ones who offered me a job still exist.

00:06:05: So, from that time on I basically changed into technology.

00:06:13: You know, my job was basically, at that time, managing digital projects, thinking about digital products and

00:06:21: you know I've had now a history of more than 20 years of building digital projects, managing them

00:06:31: And I think it was like.

00:06:34: Must have been like 7 or 8 years ago, when I decide it that I don't just want to talk about it

00:06:41: I want to touch it more, I want to be able to program, to write code myself

00:06:48: and that's also connected to the topic of artificial intelligence because when I first started to dive into the topic I was

00:06:55: So fascinated that I thought, you know, this is something, I don't just want to talk about it, I want to touch it, you know, I want to feel it and I really took a deep dive into programming.

00:07:08: And took a couple of courses. Boris Felgendreher: Formal university courses? Roland Becker: Online courses, mostly udacity

00:07:18: ok udacity courses, so courses that actually cost money

00:07:22: they were very good, I can recommend them highly, I also think they upped their game since then and it's a very good path to take.

00:07:34: after doing this I was able to build and deliver the first AI projects for our company

00:07:42: so when we started our company, JUST ADD AI, I actually wrote the code for the first couple of AI projects

00:07:51: and train neural networks. Now we have many people who are much better than me, but I'm still very technically interested

00:08:03: And I think it's also a very large benefit if you consult with the potential client

00:08:09: that you have a much deeper understanding what is actually possible and how it could be done.

00:08:18: Boris Felgendreher: Fast forward to today, how big are you today? JUST ADD AI is the place where we are recording from today, you have a full fletched

00:08:26: audio and video studio, more people left and right to us here. Roland Becker: Right now, we are 25 guys and girls

00:08:38: in the company, I think by the end of the year we will have 35, so we are on a very nice growth path

00:08:47: So at JUST ADD AI, we basically do AI-based projects for other companies

00:08:57: But we try to do that in a scalable way, so we try to look for projects that can actually

00:09:05: be turned into a product later on and be a scalable solution and then, once we have a software that is already used

00:09:14: by two or three companies, who pay us money for using that software

00:09:19: then we start to spin that software out into a separate company and scale it there, and we've already successfully done that twice, so right now we have two spin outs,

00:09:32: one is botario.

00:09:34: botario.com is a conversational AI, so we automate chats and telephony,

00:09:43: we do live chats, we do everything around conversational AI

00:09:51: And then we have scoutastic, which is a scouting software with artificial intelligence.

00:09:56: where we have a joint venture with a transfermarkt,

00:10:03: they actually acquired 50.1% of that spin-off and we are now jointly marketing that and

00:10:12: now have a good track record in the Premier League, in the German league. Boris Felgendreher: Let's dive into that, for those people who don't know: Transfermarkt is a website, that has information about players, teams... Roland Becker: And market values, it has a lot of data that others don't have, we now have the data exclusively in their platform.

00:10:28: Both of these spin-outs are now profitable, so that's very nice to see them grow, and we will spin out

00:10:38: four more companies this year, so we are on a very

00:10:41: steep growth trajectory, the next one will be lector.ai, which is the document management software, to basically classify, analyse, extract information from any sort of text or document given to other system

00:10:56: Dr. Sirko Straube: And they are all based here? Roland Becker: Yes, they are all based here. One of the companies we are going to found this year is going to be based in Hamburg actually

00:11:05: but because we have a very good team that's based in Hamburg that we want to acquire for this project so they will be based in Hamburg.

00:11:12: The other ones that are going to be built here in Bremen, at the German centre of artificial intelligence. Boris Felgendreher: Sirko, you are the scientist at this table, you are also from Berlin

00:11:27: coincidentally, what brought you here? What is your path? Dr. Sirko Straube: I started on the other side of the wall, so I'm from East Berlin

00:11:36: I left Berlin when I was 10 years old, with my mother and we went to Cologne

00:11:44: And then after finishing my school I went to Freiburg actually, to the in the very south of Germany, to study biology. So I'm no a computer scientist nor an engineer.

00:11:57: But I'm a biologist and in the later phases of the study, I started to do computer science.

00:12:06: And also taught the biologists a bit of computer science, of programming basically. Boris Felgendreher: Good choice, good forward-looking choice. Dr. Sirko Straube: Yeah I think

00:12:16: already at school I was very interested in the connection between computer science and biology

00:12:22: although I never imagined that this will take me where I am now but I was always interested in this kind of 

00:12:30: How does biology or biology consistence work and function and what actually,

00:12:37: can computer systems contribute to that model, that maybe learn from these kind of things, which is nowadays basically one of the core topics of artificial intelligence.

00:12:52: And then after that I applied for my PhD in

00:12:56: London, Tübingen and Bremen and I visited all these sites and I also decided then to come to Bremen because

00:13:09: it was a very, very inspiring group here at the university and they were very special and they were also recommended from one of  my professors in Freiburg and they said "Ok

00:13:21: go there, it's a small group", I mean they have a high reputation but I mean I applied at the Max-Planck-Institute in Tübingen so

00:13:31: at that time and still it was a very renounced institute, and worldwide known.

00:13:40: They said "Ok, it's a small group, they're doing very nice things and you can learn a bit here about" - I did human neurobiology, so studied

00:13:50: Human object recognition in the brain and

00:13:55: and then I was a year in Oldenburg after that as a postdoc and then I came to the DFKI

00:14:03: DFKI, for those who do not know: it's the German research centre for artificial intelligence.

00:14:12: Yeah it's a research centre, a research institution and dealing with all kinds of artificial intelligence and here in Bremen is a big robotics

00:14:22: institute, robotics group.

00:14:24: So I joined them and there I studied or contributed to projects dealing with men-machine-interaction, it was also interesting, not only to have a machine that

00:14:38: let's say has a dynamic or fluent interaction capabilities but won't hear.

00:14:47: People also studied "How can the human's needs for interaction be fulfilled the best by the machine".

00:14:54: And it was a time, I think maybe it was the time where the neural networks already were very popular, but we were still working with the support vector machines which is a different kind of

00:15:07: AI technology, let's the call it like that, and so that time I really worked into

00:15:16: lot about data processing, predictions, machine learning algorithms, also using other technologies of machine learning and

00:15:25: Yeah and after a while in academia, you have to decide which path to follow.

00:15:33: And at some point I decided ok,

00:15:38: maybe the view, the vision that I'm having of being a professor is maybe not my own, and I wanted to change more into the management and 

00:15:51: I also did paid courses on management and try to get more capabilities with that and

00:16:01: And then I took over a very big project, so the head of the Institute, Frank Kirchner, said "Ok.

00:16:11: We have a project where you have to coordinate 8 research department of the DFKI in Germany."

00:16:19: And so I had to do this together with a colleague from Saarbrücken and then after that

00:16:27: After having half a year of pause, my daughter was born, I came back and Frank Kirchner told me

00:16:35:

00:16:41: I had a meeting and he said "Ok.

00:16:46: We need a new deputy head and I think you are the right man for the job" and so I came into this position

00:16:59: With very high reputation, lots of duties, and it's exciting to do this job

00:17:05: Boris Felgendreher: Yeah, talk more about the DFKI, describe to us, how big is it? What's the international role, what's the presence, I mean I've

00:17:14: I've been privileged to Jana Köhler at one point in time, she was the head of the DFKI for while.

00:17:21: Fascinating story, I was amazed at how big and influential that institute is, can you share some more

00:17:28: thoughts on the importance of DFKI in the global AI world. Dr. Sirko Straube: Yeah, I think the first remarkable thing about DFKI is that it was already founded

00:17:38: 32 years ago, maybe that's something that most people don't know that,

00:17:45: already 30 years ago, people in Germany started to found the German research centre for artificial intelligence

00:17:53: and since then it has grown and grown and grown, now I think DFKI has about 1100

00:18:03: people working there and

00:18:07: It's based basically in Saarbrücken, Kaiserslautern, these are the oldest sites and Bremen is the third official site, but it's still a growing, we have

00:18:21: a place in Berlin and I think that's already very big site as well

00:18:26: although it's not an official DFKI site, but I mean that's maybe due to some some regulations and also in Lower Saxony also, it's now in Oldenburg and Osnabrück, it's spreading

00:18:40: and growing, and the idea is to have a research institute that is

00:18:45: purely devoted to doing research on artificial intelligence.

00:18:50: With all kind of facettes. I mean, for example, we have a group in Berlin that is dealing with interactive textile, so they are trying to

00:19:05: do research a new textile technology where you can somehow monitor human body parameters.

00:19:13: We have the robotics group here in Bremen, I think it's still the biggest research department of the DFKI with

00:19:21: about 100 scientists working here plus

00:19:26: around about 50 students workers and in Bremen, we also have a second

00:19:32: research department with Rolf Drechsler, we just had him here in the podcast, which is called cyber-physical systems,

00:19:41: and it's also about 50 people working there, so we have more than 200 people

00:19:47: working here at DFKI just in Bremen. Boris Felgendreher: What was the original rational for planting that beachhead here in Bremen?

00:19:56: Was it the connection to the university or the industry or? Dr. Sirko Straube: So, there was always the connection with the university. Roland Becker: We actually talked about that in the podcast with Frank Kirchner.

00:20:10: And as far as I remember, he pointed to some political decisions, which were taken in Bremen a little over 20 years ago.

00:20:23: Funding for the University,

00:20:25: funding also for that focus on artificial intelligence and they actually were able to design a package,

00:20:34: that was also interesting enough for Frank Kirchner, 

00:20:37: who was at that time a leading roboticist in Boston. It was interesting enough for him to lure him back here to Bremen

00:20:50: and because the opportunity looked brilliant, so I think it's some very important, very good political decisions,

00:20:56: which are mostly long-term funding decisions were taken at the time

00:21:03: for the university and then for the DFKI. Dr. Sirko Straube: You have now described what followed then - I mean there was that package - and I mean one part of DFKI is that the university and DFKI are always connected because the department heads are also

00:21:20: they are university professors and have their university group. We also have a big group at the University of Bremen that is also headed by Frank.

00:21:31: And, but I don't know actually, I think someone had to ask him and so I mean

00:21:37: Because the Federal State of Bremen is also,

00:21:43: sitting in the shareholder board of DFKI, so the three federal states are also part of the shareholder board and there is to be some decision and actually I don't know,

00:21:53: who was the first who took this decision but what came out was

00:22:00: was that a political decision was made and the bundle was

00:22:06: packed and this worked - yes. Boris Felgendreher: On the topic of shareholders, Google is an investor?

00:22:14: Dr. Sirko Straube: Yeah, Google was the first investor, or the only investor in European research, was at DFKI. Boris Felgendreher: What was the rational there? What are they in for it?

00:22:24: Dr. Sirko Straube: Yeah I mean the investors have a special role at DFKI. So there's a shareholder board and we have Airbus, Google, we have a lot of big companies in the shareholder board.

00:22:36: It's more a strategical decision that these companies go in the shareholder board, because they

00:22:44: they are not buying parts of DFKI, but they just

00:22:52: form important strategical decisions in this board and they also commit a certain level to DFKI and say "Some of our research has been done with DFKI" and if they

00:23:06: let's say qualify through this process then they become a member of the shareholder board, so there is a certain number of places

00:23:16: available. Roland Becker: It's a very exclusive club. As DFKI it's not for profit, so they are certainly not in it because of profit.

00:23:26: There must be something else that is inside information obviously. Dr. Sirko Straube: Yeah I think it's information and it's also, DFKI's concept is really as a public-private partnership.

00:23:39: So it's really that you have the BMBF, the German ministry for research and education.

00:23:47: You have the federal states and you have all the companies there and they're also together. I think that's also something that is very attractive for these companies to have an exchange about artificial intelligence on this level.

00:24:01: Boris Felgendreher: Yeah, and you touch a little bit on different parts of the world. You mentioned Boston for example as sort of a hotbed for talent and for a sort of 

00:24:10: an ecosystem of AI for example. How does Bremen compare to other sort of hotbeds and ecosystems around the world? Like where are? How do we compare in Bremen?

00:24:22: Both: In terms of AI?

00:24:25: Dr. Sirko Straube: One thing that is still very remarkable about DFKI is that it is a research centre, only a public research centre, only focusing on artificial intelligence.

00:24:38: And I think with that

00:24:39: focus, we have been many, many years the only ones and also the biggest, let's say, research centre and

00:24:48: to my knowledge maybe we still are but, I mean there's a big but, because many companies invest

00:24:57: Really high amounts of money into this field. I mean Toyota started, we have lots of companies in the Silicon Valley, all over the place so I mean

00:25:07: already 10 years ago or 20 years ago a run has started on this technology, so I think concerning research we are still

00:25:17: let's say, we still have the leading role worldwide. Many of our researchers, also from the universities, go to the US or also to China more and more

00:25:30: but

00:25:32: yeah I think the hard part is also the transfer of this knowledge and of this high competence to the industry. I think that's really

00:25:42: That's a hard part here. Roland Becker: I think that is something that is now getting better but as you asked I think Bremen has actually influenced also

00:25:52: many of the large companies that that the people know. For example I know that one of the first employees of SpaceX is actually from Bremen, one of the core developers from Siri

00:26:02: actually from Bremen. So many of the leading and known products were actually designed, also co-designed by people

00:26:11: who actually learned what they're good at here in Bremen and this is something that I think is not widely known.

00:26:20: I think in the science part Bremen is regarded quite highly, especially in areas like robotics.

00:26:30: There was just recently this international technology ranking which are the hottest professors and people to watch,

00:26:38: and I think, for example, Professor Michael Beetz from the International Institute for artificial intelligence here was worldwide on place number four.

00:26:50: So he is just one, one of the many very excellent people in the field of research.

00:26:59: What Bremen has not been very good at in the past,

00:27:03: actually keeping these people in Bremen, keeping all these brilliant talents who come out of

00:27:10: the classes of these brilliant professors here in Bremen and that's what we are now actually trying to do more here, so to really convert this talent into business opportunities here

00:27:22: in Bremen and I think if, for example, if I look at it as an entrepreneur

00:27:30: I look at opportunities, what are cool things that you could do here in Bremen, I think building AI companies or AI departments for larger companies

00:27:42: here in Bremen is a very large opportunity because we have so much talent, you know, coming out of the universities, out of all the different institutes we have here.

00:27:52: That it's really no problem to find really highly qualified people here

00:27:58: so, human resources are not a limiting growth factor here. Boris Felgendreher: And it's amazing, because it's not a big city, we are not talking about London size or Paris size, we are talking about a city that's 500.000 people

00:28:12: so it's a high ratio of talent to regular population. Roland Becker: I think in some of my German podcast and talks

00:28:22: I tend to say that I think the ratio for AI talent to the general population

00:28:28: it's probably highest here. I mean prove me wrong, you know. Of course, in total there may be more AI researchers in Munich or Berlin, whatever.

00:28:41: But you know, in ratio to the general population, I think here in Bremen, at least in Germany, maybe even in Europe, you have the highest density.

00:28:48: Boris Felgendreher: And how long has it been so? I mean has that just recently developed? Has it been like that for the last 10 years or is it just something that recently developed?

00:28:57:

00:28:59: Dr. Sirko Straube: I think it's a recently developing and more and more influencing thing, so I mean the robotics innovation centre where I work in

00:29:09: we started 12 years ago

00:29:11: and it takes some time and also, I mean one thing the DFKI tries to do is also to support people who

00:29:20: if they want to do spin-offs and these kinds of things.

00:29:24: And this also in Bremen now, I mean with robotics one has to say it's a hard topic 

00:29:31: to found your own company because I mean it's not only software but you also need the hardware parts so it's

00:29:38: usually a large investment, maybe it's getting less and less, because you have more commercial products that you can live with.

00:29:47: But I think that it is really starting at the moment or has already started, so we have of course the company from Roland, JAAI, and all the other companies here and they're growing and we have

00:30:00: Also two robotics companies that came to Bremen that have, let's say, still small seeds here but they are also growing and I think that will be more in the future, so I think

00:30:14: And we have all these, you know, we have Airbus here, we have a lot of companies here that are also - so we have the big companies, we have the in-between companies, we have

00:30:24: the spin-offs. Some spin-offs also fail, I mean, we have also had other companies that, of course they started and then they failed

00:30:35: but I think there is a kind of the ecosystem slowly developing here and Bremen and, as Roland said, we have a very high ratio of people.

00:30:45: If you look at the population and then if you look at the people that are really very well educated in this kind of technology development.

00:30:55: Roland Becker: If you think about it, so I'm talking about the the science part and the Education part of it so,

00:31:05: many people are coming to Bremen to study artificial intelligence, robotics and stuff like that, from all of Germany but also from abroad you know, like many people come specifically from abroad here to Bremen

00:31:19: to study specifically here because they know they're going to get a super great education, they are going to work

00:31:25: with the best in the field internationally, globally. Boris Felgendreher: And it's more affordable than MIT. Roland Becker: Totally.

00:31:35: And then what we are getting now better, this funnel is actually accelerating, I also see the same way, it has not been like this for the last 20-years now.

00:31:44: But we have seen I think especially in the last like 5 years of constant acceleration.

00:31:51: Of that cycle, of you know, people being drawn to Bremen to actually build the expertise.

00:31:58: And then, you know, coming out of the universities, out of all the different institutes.

00:32:03: And then, yeah, basically either staying here - what is also interesting is, of course, most of these people could

00:32:11: directly go abroad to the value or to whatever and make big money.

00:32:16: But for most of the young people today making big money is not their first concern.

00:32:21: Most of the young people that I talk to and that I know say "You know what. I actually, I love Bremen and I would love to stay here and you know

00:32:29: I'm just, what is just important for me is not so much the money, I want to have an interesting job.

00:32:35: I want to continue to learn, I want to learn from people who have more experience in the in the actual,

00:32:44: in actually deploying AI into practice", you know, and

00:32:50: I think that's one of the reasons why in the past we have lost many of these talents to other cities, other countries and now there are more and more opportunities for these people

00:33:04: directly here in Bremen, companies like us, you mention a couple of other companies where they can find a home,

00:33:10: where they can build on the expertise, you know, continue to learn do something cool and then, you know, making more money is, like I said, it's not the main reason, it's one of the major things for people.

00:33:23: Like, you know, when we talk about investing more money, maybe even into universities and research facilities.

00:33:30: Then people tend to say "Well but then, you know, we fund all this education and then the people go abroad and make the money there and we never see them back",

00:33:39: it's just not true. Of course it happens but these people want to stay here because it's a beautiful city

00:33:49: just look outside, nice sunny day at the harbour. Dr Sirko Straube: I absolutely can confirm this, I think as well

00:33:58: one thing that is maybe underestimated in Bremen, is that the city has a very

00:34:04: let's say high life quality for people that come here and that come you from abroad, that come from all over the world and there are, I would absolutely agree, many people come here and they really

00:34:18: Find home here or like to stay here at least for longer time and I mean

00:34:24: it was also the same with me. When I came to Bremen I also thought, ok this might be some place where I could stay and I stayed here and I really like the city and

00:34:37: I think that's something that is a bit exceptional and sometimes I have the feeling that some people in Bremen,

00:34:44: of course people in Bremen love their City but maybe

00:34:48: some people do not see, the let's say special beauty of the city that is in contrast to many other cities. It's a hidden gem. Roland Becker: Is Bremen Germanys' Austin?

00:35:00: Boris Felgendreher: I lived in Austin Texas, so I should know. Roland Becker: I just heard all the rumours like people are leaving California and Silicon Valley, everybody seems to I don't know

00:35:12: Boris Felgendreher: Think about Austin as the liberal enclave

00:35:16: inside the conservative Texas, lots of students, lots of artist, lots of live music, tech companies.

00:35:22: And as you said now there's a lot of movement going on between Silicon Valley.

00:35:27: People going to Austin Texas, also to Miami, it's another hotspot right now that never used to be on people's radars when it comes to tech startups and all the sudden the crypto scene is moving over there so

00:35:36: There's movement, so it's not located anymore to like small number of spots in the world but it's spreading out.

00:35:43: Maybe Bremen is benefiting from that. Dr. Sirko Straube: The German Miami might be stretched too far. Roland Becker: I said Austin but Miami, why not?

00:35:52: We are living close to the sea, we have the harbour. Boris Felgendreher: I lived in both places and now I live back in Bremen,

00:36:02: the rest is history, I let you fill in the gaps

00:36:08: Roland, talk to me about the couple of specific projects, you mentioned a couple of use cases you had earlier but

00:36:16: talk to me about some really exciting new use cases coming out of your shop right now that are

00:36:20: very fresh of the press so to speak. Roland Becker: So in general, so yeah, very fresh of the press that's of course always with us in economic sense,

00:36:32: because we don't do research really, we try to follow the research very hard,

00:36:41: to really know about all the latest ideas and innovations.

00:36:46: That have been found and then we try to filter

00:36:51: which one of these can we actually use in practice right? And then actually deploy that into production with a very short cycle, so that we always have the latest state-of-the-art AI technology,

00:37:03: deployed in our projects, that's what we do.

00:37:06: And then of course, what I could talk about for example, there's actually a couple of interesting projects that we are currently delivering.

00:37:16: So

00:37:17: one group of projects is actually focusing around our botario.com conversation-AI software, where we now see, first of all a large

00:37:30: wave of adoption for chatbots and livechats, because people, especially since the pandemic

00:37:37: are beginning to understand that much of their communication with consumers or clients is routine communication.

00:37:46: And they could actually deliver that at high quality 24/7 using a very well designed bot.

00:37:53: And then, of course, you can not just do that by chatting to a bot on a website,

00:37:58: but you can also use basically the same AI behind it to automate your telephony.

00:38:05: So we are building a trilingual phone bots

00:38:11: for a company in Luxembourg, which is actually able to negotiate meeting dates

00:38:20: for different locations with people on the phone and the bot at the company, so quite a complex thing: When are you free , when do we have free slots?

00:38:34: No I'm usually free Wednesdays in the afternoon and Thursday in the morning, with Thursday next week, stuff like that so

00:38:43: when you talk to these kinds of bots, it very much feels next-level.

00:38:49: And it's not just, the thing about all the products we do is we're not building Showcases, we're just building

00:38:57: really production-ready services, that are in reality delivering value, you know, and not just some demo cases like Google did very impressively, couple of years ago

00:39:11: with their automatic phone calls, you might remember that.

00:39:16: Can't even remember the project name anymore but anyway, so we're doing this for real today.

00:39:25: That's very interesting, another very interesting project we're working on, which has just been publicly announced.

00:39:32: Is together with the Bremen tram company, so the BSAG

00:39:39: Boris Felgendreher: You didn't say Trump, you said tram. Roland Becker: I said tram, tram, tram, tram, the public transport.

00:39:50: So we have these trains that drive all around Bremen and,

00:39:58: none of these trains has any assistance

00:40:01: system right? So every modern car has some sort of assistance system which warns you when

00:40:08: people are on the road or, you know. Boris Felgendreher: So it's for the tram drivers? Roland Becker: Yes, it's actually for the tram drivers. So we are building retrofit system that is inspired by comma.ai

00:40:22: George Hotz, if you heard of him, and his company is doing awesome work.

00:40:27: And so they're just taking little phone-like devices and putting them on the windshield of cars.

00:40:34: Connecting that with OBCD-plugs into the car and then you have level 2 autonomy. That's super cool

00:40:44: and we thought, you know, why can't you do the same for a tram because it should be easier because

00:40:50: this is many less levels of freedom for a tram, you know because it's driving on railway, so yeah and the thing is

00:41:01: Bremen just aquired new trams from Siemens

00:41:06: for somewhat over 400 million and these actually do have the sensors, cameras and Lidar and that sort of stuff.

00:41:16: But no working assistance system yet and but all of the old trams still have to be used for at least 15 years.

00:41:26: So and they have nothing, you know, so we thought why not build an assistance system that we can just put into the old trams.

00:41:34: And which can then assist the drivers to tell them about speed limit which they haven't seen, tell them about people on the railway.

00:41:44: And stuff like that. I think that's all so very, very exciting. Boris Felgendreher: Yeah, are all of your clients from Germany or are you also selling to clients internationally?

00:41:53: Roland Becker: So actually, right now most of our clients are here in Germany, that's just because there are so many opportunities here. Actually we are not doing any sells.

00:42:05: We currently rely on people calling us or contacting us "We heard of your company, we need your services", so it's really like that, we also

00:42:15: decline projects if we think it's no strategic fit for us.

00:42:19: So and it's a very lucky position obviously to be in, so we are just picking the lowest hanging fruits currently.

00:42:28: I think there are many, many opportunities globally.

00:42:32: And we are currently looking into different ideas how we could expand our business.

00:42:39: Also internationally for our spin-offs of course. We have like, for example, the Scoutastics software like I said,

00:42:50: we are in the UK very successful and also in other European markets. We are also making offers, moves in the Americas, so USA but also South America,

00:43:04: so that's already happening

00:43:06: Boris Felgendreher: Do you have plans to open offices globally or what is the ultimate plan? Roland Becker: Yeah maybe, so far it's this Bremen office, so the next office we will open will be in Hamburg this year

00:43:21: And probably we could also need more offices here in Germany.

00:43:27: Like I said we are a very dynamically growing.

00:43:33: startup and we have doubled our revenues and margin last year and we want

00:43:40: to continue to grow at least at that pace, yeah like I said I can't talk too much about it but we're currently also looking at opportunities

00:43:50: to scale our business also with external growth, so far

00:43:56: we've only been growing organically, so just lived from the money we made and reinvested that and we still can keep that quite high pace of growth dynamic without external money.

00:44:11: Who knows what the future brings

00:44:13: Boris Felgendreher: Sirko, you probably also have a pretty good overview of the different startups that come out of Bremen, any good examples of companies that made it past the boundaries of Bremen and then Germany and growing internationally.

00:44:26: Roland Becker: I know one, so that the latest and biggest recent success is obviously Ubimax

00:44:32: Boris Felgendreher: Also right across from here right? Roland Becker: That's also one of the reasons we've moved here,

00:44:38: we can always look at Ubimax who already did it, you know. Dr. Sirko Straube: I think Ubimax came out of the BIBA, it's the Bremen logistics.

00:44:51: research centre, so to say.

00:44:53: Yes, so this was a very successful thing. I think there is another one, a small company that's called cellumation, they have omni-directional. Boris Felgendreher: Hendrik Thamer, the CEO of cellumation was on 

00:45:07: one of the "Go Global! Bremen Business Talks", very exciting the sort of conveyor belt that moves,

00:45:13: almost looks like it's Tetris right where it's magically moving around and it's using artificial intelligence to some extent right?

00:45:21: Roland Becker: If you just quickly come back to Ubimax, they actually sold their company to TeamViewer for 135 million Euros I think, one of the one of the biggest exits

00:45:32: in the last year for sure, very successful. Dr. Sirko Straube: I think one of the company's that I was mentioning but not naming just before was, so we have for example a group of

00:45:47: Kraken which is a Canadian

00:45:51: underwater company that came here, they hired some of us people

00:45:59: that worked in our institute and they founded a small company called Kraken Robotics here.

00:46:07: And also another company is ROSEN, we also had Peter Kampmann here in our "THINK REACTOR" podcast

00:46:17: so they are doing inspection of pipelines, mostly from the inside

00:46:24: but they also said "Let's have a team here in Bremen now" and they're also growing, that is also

00:46:31: looking at new technologies from underwater robotics, so I think, as I said there's a lot of things

00:46:37: happening at the moment and I would say if we

00:46:41: would talk again 10 years from now then it would be completely different story. Boris Felgendreher: Which is probably a good transition. I do want to talk to you about the future and some predictions, Roland why

00:46:52: don't you jump in? Roland Becker: So I'm also a board member of bremen digitalmedia which is the local digital

00:47:02: industry cluster, where we also have our AI cluster, Bremen.AI and we recently

00:47:11: did a study regarding AI adoption in companies in Bremen.

00:47:17: And we actually found that the adoption, so the rate of companies that actually are using AI or developing AI,

00:47:26: is at least double or 3 times higher than in Germany in general, so that was very impressive.

00:47:38: Boris Felgendreher: What do you contribute that to? What's your best guess? Roland Becker: My best guess is that high density of AI Talent that we have in, you know in comparison to the general population.

00:47:49: Which, you know the cause for that we already discussed

00:47:54: Boris Felgendreher: But that wouldn't necessarily have to translate into companies in Bremen actually using it right? Roland Becker: Eventually, eventually yes I think.

00:48:02: Dr. Sirko Straube: And I mean, we have for example the neusta company here, we have a lot of IT companies and I think, I would guess that they're also using lots of

00:48:13: AI technology inside there companies. Roland Becker: They have their own AI department for sure, team neusta was ranked number one internet agency in Germany last year or the year before.

00:48:27: And they have 1000+ employees also, so quite larger player, we also have

00:48:33: hmmh here in Bremen, MERENTIS and many large digital companies, who of course also

00:48:43: contribute a quite large number to the workforce in Bremen and I think

00:48:49: also in the larger demographic setting, you can see that, you know because Bremen was traditionally all about the harbour

00:48:57: then was logistics and space also,

00:49:03: and you can see that more and more the IT business is taking over

00:49:07: Boris Felgendreher: Yeah what about AI usage in those sort of Old Guard traditional industries like harbours, transport, logistics

00:49:15: in food and beverage, in that sort of bread and butter of Bremen industry? Roland Becker: So they are actually awake, so now I must be a little bit careful because we work for some of those companies but we must not say it publicly.

00:49:28: What I can tell you, maybe as a general picture and we do not work for all these companies which I mention, so that makes it a little bit anonymous

00:49:39: large players, like in logistics, of course you have BLG and they are already working on a AI projects.

00:49:48: We have large players like Cordes & Graefe which is one of the largest companies here in the Bremen area.

00:49:56: 6 billion Euros turnover, they have a team working on AI. Airbus, of course they work on AI. I think all of the really big players

00:50:09: are already open-minded and are already starting to explore that topic. It has not been for too long that they have started that to to be fair.

00:50:18: I think like 5 years ago many of these

00:50:21: didn't have any activities in that area but maybe also with the Bremen.AI we contribute

00:50:28: to them being more active in in that area and I think now I really can't think of any major company that you would know that,

00:50:38: who would have really no activity at all in that area. Dr. Sirko Straube: If I remember maybe 7 or 8 years ago already I started to give presentations for

00:50:51: many of the companies that Roland was mentioning and many others 

00:50:56: and they were all very interested but I would also agree that they didn't have many activities there and I think

00:51:03: for example, I mean I can not speak for a whole field but my impression was always that for example the logistics

00:51:11: companies, that they were very much driven by

00:51:16: the current market situation they were in and they were driven by the contracts, they had the logistic contract, they're also limited in time.

00:51:25: There was, let's say,

00:51:29: there was very little space for long-term changes of these companies, so I think that was why they had their,

00:51:37: that was their biggest hurdle they had to overcome, so to say,

00:51:44: but I think that has changed a bit and maybe also in Bremen there starts, maybe also due to Bremen.AI, due to DFKI, due to all these,

00:51:55: these names and people around that now, and Bremen always has this very high advantage that

00:52:03: people know each other, so people know each other on the political side, on the research side, on the business side.

00:52:11: And with Bremen AI, I think people came even closer together and maybe there's a certain common

00:52:20: growing consciousness for this. Boris Felgendreher: Sort of a close-knit community of business and AI in Bremen? Roland Becker: Yes, in general we have a saying in Bremen which is "The ways in Bremen are short"

00:52:34: because as you said it's a geographically quite small area, very densely packed with talent and also it's very interesting larger companies and industries.

00:52:46: Because of the harbour obviously that we have and the big history that Bremen has.

00:52:53: And I think that it starts to come together now. Boris Felgendreher: The other company that we haven't talked about is Mercedes Benz, one of the largest companies worldwide here in Bremen, are they doing stuff with AI in production?

00:53:06: Here? I mean I'm not talking about autonomous cars which they probably do somewhere else, in Sindelfingen or Stuttgart or wherever. What do they do here, you know?

00:53:13: Dr. Sirko Straube: So we had several meetings and workshops with them, I think they are very much centralised somehow, so many decisions are maybe not made

00:53:26: here, as far as I can tell, though I'm not sure what they do actually here but.

00:53:33: Roland Becker: I'm also not sure if I could talk about that. Boris Felgendreher: Knowledge you can't share. Dr. Sirko Straube: One thing I wanted to say is that I think many of these old

00:53:47: or maybe very successful branches, so to say, 

00:53:57: they really have a difficulties to change, so I remember for example when I was 10 years ago I think I was in a

00:54:03: meeting with a DLR, the German

00:54:08: Space agency so to say, it was not here, it was somewhere else

00:54:18: We presented our technology, the robots, autonomous exploration is a very important field and we have to invest in that and have to do research on that and one of the guys from the DLR said

00:54:31: This will never be used in space and I mean it was one of the first meetings I had when I came to DFKI and I was really shocked and I thought

00:54:42: Ok if we don't need autonomous robots in space, where else? And I think what he meant at that time was that it's expensive to fly these things, if we have an autonomy function in this robot

00:54:57: that is failing, it costs millions and billions of euros or dollars and I mean, of course

00:55:05: this is also changing now and we have also robots on Mars for example, but they are also still, let's say

00:55:13: very semi-autonomous. I think in the media there is a bit more autonomy suggested than they really have.

00:55:22: But I think that's maybe one example where it is clear that this is coming. Autonomous robots will be used in the field and I think autonomous logistics

00:55:32: are already used in many fields, if you look for example at Amazon or so.

00:55:39: But if you have a good business running and the processes are fixed then it's really hard to change all of that and AI is all about changing things, so you have to really rethink all your processes

00:55:52: from scratch so to say, maybe not

00:55:55: completely from scratch but a lot of things have to really be redefined and thought completely new and I think

00:56:03: that's the hard task of these companies. Boris Felgendreher: Let's now move to the promised future outlook. I'd love your guys's take on the next 5-years, what you see, maybe also in regards in connection to Bremen, how this ecosystem of AI

00:56:17: in Bremen

00:56:18: will idealy evolve over the next 5 or 10 years? Maybe if you had a magic wand, what would you wish would happen in next 5 to 10 years? Roland Becker: What I see, I don't need no magic wand for that, I think

00:56:33: this is a kind of explode really, because what we're seeing right now is exponential growth.

00:56:41: Like Sirko just said, like 5, 6 or 7 years ago

00:56:46: there was hardly any activity in the economy really, we had only science activity.

00:56:52: And now this starts to really

00:56:56: move into the economy part, we have more and more startup companies coming up, more of the older larger companies adopting the technology.

00:57:06: And as I said earlier we do have this large inflow of AI talent here.

00:57:14: And we're starting more and more to actually being able to absorb that,

00:57:19: and put that into practice really and yeah that's why I'm convinced that Bremen is actually one of the super hot locations

00:57:31: For AI in Germany and, you know, people will start to recognise it in the coming years.

00:57:39: Boris Felgendreher: Sirko, anything to add there? Dr. Sirko Straube: Yeah, what should you add there? Roland Becker: You could add the scientific part. Dr. Sirko Straube: I'm always a bit sceptical if every growth is exponential because now we talk about

00:57:52: exponential growth about everything. Boris Felgendreher: Little inflationary right, that term? Dr. Sirko Straube: Yeah but

00:57:59: I mean with the rest that Roland said I would completely agree.

00:58:04: I think also from the research part, I mean we have now just made a new building or a second phase of our building, so also the infrastructure from DFKI but also from the University is going there, so I think there will be more talents.

00:58:19: I think also there will be lots more companies that come and also DFKI, so now we have this small

00:58:28: company that is called DFKI Industrials and we're trying to somehow

00:58:32: transfer our technology or find better channels to transfer the technologies to companies. I think in Germany it's not so easy.

00:58:44: But maybe this also has its justification, but

00:58:47: it's, at DFKI we're doing research that should be applied research and that should also be transferred to the companies.

00:58:59: And I think more and more of this technology will also reach the market and in some form of the other and I think lots of this will happen in Bremen so

00:59:10: Yeah, I think with everything we've said,

00:59:13: maybe it will explode but it will definitely grow I think. Boris Felgendreher: Yeah, fingers crossed, awesome, exciting times.

00:59:22: Roland, Sirko, thank you very much for todays conversation, I had a blast, good stuff.

00:59:28: Look at my watch. I don't know how long we went. I think longer than our typical podcast but not as long as yours are typically right,

00:59:34: your's are typically two to three hours. Roland Becker: Yeah, we have no time constraints but we knew the rules would be different today, thanks for having us. We should do this more often. Dr. Sirko Straube: Really, really nice and I think

00:59:49: let's have more of that, thank you.

00:59:54: Boris Felgendreher: Alright that was the "Go Global! Bremen Business Talks" episode about Artificial Intelligence made in Bremen.

01:00:02: Hope you enjoyed todays show, if so please subscribe to the podcast so you don't miss any of the future episodes.

01:00:08: We have a lot of interesting guests and topics about the businesses, the technologies and the people that connect the cities of Bremen and Bremerhaven into the rest of the world coming up.

01:00:17: Music.