"Just give it a try!" TUM alumnus Florian Grigoleit on his learnings from the SkyDeck Accelerator Program
Florian Grigoleit is the CEO and co-founder of modelwise, a TUM startup that provides AI-based software for automated safety analysis. Florian received his master’s and PhD degrees in computer science and AI-based modeling at TUM, where he met his co-founders. Encouraged by UC Berkeley, he applied for the prestigious SkyDeck Accelerator Program with modelwise – and was successful. TUM San Francisco liaison officer Jeff Ouimet met with the TUM alumnus to talk about his journey from student to entrepreneur and CEO.
Florian, you went from a TUM PhD candidate to becoming the successful CEO and co-founder of a TUM spinoff. How did you do that?
I studied mechanical engineering for my bachelor’s degree, but I became more interested in computer science. I received a Fulbright scholarship to study at Washington State University, where I worked in computer engineering. I then came back and started my PhD at TUM in the computer science department, where I worked in model-based reasoning and AI. In 2016, I got another grant for Software Campus, which is a German initiative to educate IT people in management and entrepreneurship. During this project I met Arnold Bitner and Iliya Valchev, who are my co-founders now. After this project, we determined that the technology was worth pursuing so we applied for an Exist research grant. We got it, and eventually spun out of TUM last summer. We were then approached by a scout from the SkyDeck Accelerator Program. We applied, and were accepted.
When did you start thinking of yourself as an entrepreneur?
Very recently, actually. I still think of myself as an engineer but I can’t tell investors that! Actually, it was during my time at SkyDeck. That really changed my perspective. During Exist, there was still a lot of research involved, and it wasn’t too much different from my PhD time. I noticed that I have a strong sense for strategy and very high level thinking and decision making. That is not the best for an engineer, but is quite helpful for a CEO.
Did studying at TUM open up opportunities you would not have had at another university?
Yes. In some ways I was really lucky with the research group I got into. I was actually working for two different research groups: one for my PhD and another for my project. I had two really good advisors who gave me a broader perspective than just research. When I was at Washington State University there were two PhD students who had a startup. That was one of my first contacts with people who had a technology and were trying to build a company. I hadn’t actually experienced that in Munich before, and I found it really exciting.
Did studying and spending time in North America shape your sense of entrepreneurship? Would you recommend spending time abroad to budding entrepreneurs at TUM?
I’ve been abroad a couple of times: once in France during my bachelor’s, a year in the US during my master’s, a couple of months in Australia during my PhD, and now just recently back in the States at SkyDeck. I think the specific experience is one thing, but you are confronted with so many different challenges, unknowns, and you gain a lot of confidence, and that’s important for entrepreneurship. You are faced with so many different topics that you’ve never encountered before. Like at the moment, I’m talking a lot with a US tax consultant – something I never imagined I would do in my life.
How would you describe your company modelwise?
Modelwise is, at its core, an engineering company. What we have is a technology for creating and analyzing, or reasoning using models of technical systems. At the moment, it is mainly electronics or integrated circuits. That is what we do from research. What we are monetizing at the moment is the analytical part. We can analyze whether the technical systems comply with requirements or safety standards.
For example, about three years ago two Boeing planes crashed. There was a small error in a sensor that cascaded through the autopilot system and caused the deaths of 400 or so people. With our software, this fault would have been detected. This is our first product on the market. What we are building now is a data application, or platform, which we integrate into the entire value chain of electronics, from the chip manufacturer to the end producer. The product is called Paitron.
Our core technology revolves around models. We have a dozen applications in mind but the core technology is abstracting and representing technology. For example, with electronic designs you can make a lot of very fast analyses and diagnostics. Safety analysis, reliability studies – there are tons of applications for it.
How was modelwise created?
It actually started in our research group back in 2018. My academic advisor had a meeting with one of my co-founders, and they asked me to join. They each thought the other one had filled me in, so I walked into the meeting totally blind! I quickly realized they wanted to start a company, and I already knew I wanted to start one. We asked another student if he wanted to join and, since we had complementary skillsets and could identify a big opportunity, we decided to form the company.
How does your product Paitron change the world?
Essentially, what we do is make electronic designs. Let’s take a step back: hardware design processes are about 20 years behind software design processes. Hardware design involves a lot of documentation but very little testing and experimentation. What we do is enable agile hardware design. Our customers could be anyone working with electronics or technology in the future, but at the moment we are focusing on customers in industrial automation and process automation. For example, chemical plants, the petrochemical industry, companies that produce sensors, electronics, control units, and those kinds of things. Our technology can apply to nearly any process, independent of the hardware. It doesn’t matter if the electronics are for a plane, a car, some machinery or even your smartwatch.
Are there other markets you want modelwise to tap in the future? Do you have international expansion plans?
The first step is getting into every electronic domain: aeronautics, automotive, you name it. It’s a five to seven-year mission to get into all of these markets. After that, perhaps semiconductors. At that point, we will already be a multi-billion-dollar business. We already have a client in France and one in Japan, and soon the first in the US. By the nature of our business, we must be international.
How did modelwise become part of the UC Berkeley SkyDeck accelerator program?
I had actually heard about SkyDeck, but I thought it was just a Silicon Valley thing. Then, a scout from SkyDeck approached me and encouraged us to apply. The application process was not very difficult. We had to fill out an online form, submit a pitch deck and a one-minute video presentation. That only took a day or two. Then there were three interviews: the first was just about cultural fit and getting to know the team. But then, there was a first pitch in front of a jury – and that was the point where we thought: “OK, we’re out.” They grilled us brutally! We learned later that this is intentional, to weed out everyone who is not up to their standards. But it turned out okay, and we had a final round with an even larger jury. This is much nicer because the goal is to figure out how they can help you, and whether they have the expertise and resources to really make an impact on your business. But in the end we got accepted. My co-founder Arnold Bitner and I are the ones who actually went to Berkeley.
What was the atmosphere at SkyDeck like? How important is being there?
The first two months we focused on finding mentors and attending workshops. There were so many workshops! About 50 or 60 of them covering everything you can think of about startups. The next two months were focused on business modeling, followed by a final two months we spent on fundraising.
The first thing I noticed was this American Silicon Valley thing. At the kickoff it started with “Ok, we grilled you brutally but now we’re on your side, and we want to see you reach the moon. We don’t work small. Everyone here has the potential to become a billion-dollar business and that’s what we’re aiming for – to change the world.” Then we had a meeting with Chon Tang from the SkyDeck Fund. It blew my mind. We had a 45-minute meeting and went through everything: where we are with finance, with business development, with products, with market analysis, with fundraising. He was super to the point, and he really spotted our weak points. Mr Tang is brilliant and very experienced. He gave us amazing contacts – we left with six or seven introductions to potential mentors and customers.
In my first week at SkyDeck, I got into the elevator with an elderly gentleman. I introduced myself and we had a 13-floor elevator talk. It turned out he’s a business angel and SkyDeck advisor. We exchanged LinkedIn profiles, met for coffee, and he ended up introducing me to his angel network. He helped me with our pitch deck and met with me four or five times to advise us on our business plan and tell us what investors are looking for. I made several good contacts like this in Silicon Valley. It’s a really open atmosphere. People are interested in what you’re doing, and at SkyDeck people are really keen on helping. With TUM’s great alumni network, we could do better to recreate this atmosphere in Munich. For example, to provide more real-world consulting and advising.
What did you learn at SkyDeck that you could not have learned anywhere else?
Maybe “anywhere else” isn’t the right term, but what I found unique was learning to think like an entrepreneur. Not like engineers trying to sell a product or researcher who have a cool technology. Instead, I learned to take a real business perspective. It’s a fast-paced environment. But there is also risk tolerance. For example, when SkyDeck forms a cohort, they also include a “jockey race.” It’s for a group that may not have a great business model at the moment, but the SkyDeck team feels that there is something so amazing about the team or one member that they will succeed even if they throw the technology overboard. They take the risk and let them into the program. That’s not something that would happen in the Munich ecosystem.
What advice would you give to budding entrepreneurs at TUM?
Two things. First, just give it a try. Second, talk to other founders and get their experience. Ask really pointed questions, like How did you handle IP contracts with TUM? How did you handle your first sale? How did you get to fundraising? That’s something we didn’t do in the beginning and now I see it as a huge difference. Here, we all have to discover the same thing over and over because there isn’t that much interaction between the teams. At SkyDeck, if I had a question I would just pop it in a Slack channel and I had four or five answers within an hour.
Thank you very much for the interview, Florian. And good luck for the future!