Big data gets smart

SiaSearch (now Scale Nucleus) is one of the first ventures created at Merantix. They dub themselves as the “smart data refinery”. They are developing an intelligent raw-data processing framework, something akin to Google Photos for data. This technology is enabling companies to create large scale AI solutions by making data manipulation easier and more intuitive.

The company was acquired in 2021 by ScaleAI, Silicon Valley’s leading data management company valued at $7.3 bn.

Current Funding
Team Members
Lorem ipsum tbd.

The Founders

Mark Pfeiffer
Co-Founder & CTO

Coming from a strong engineering background at ETH Zurich and Berkley, Mark has been a member of the board at HackZurich and a research assistant at ETH.

Clemens Viernickel
Co-Founder & CEO

Clemens' career has always been a combination of business and tech: Having studied at the University of St. Gallen and then moved to Google in strategy positions.

your Venture

We are looking for ambitious and driven talents, who are ready to start ideating and building with us. We have created a five step process that will orchestrate you to creating an AI venture with lasting impact.

our Ventures

We have already built successful ventures in various industries, all using the potentials of AI to create a better future for society. From joining a venture pre-incubation, to being part of a hypergrowth phase, find your calling here.


After working for Google’s strategy branch in London, Clemens joined Merantix in 2018 in an operational role. Alongside our founders, Rasmus and Adrian, he supported the initial Merantix fundraising efforts and the overall Venture Studio setup. One of the first projects during this time was in the automotive industry - where he decided to found his own company, SiaSearch. Having completed prestigious research projects and signed first customers, SiaSearch has started to apply its technology stack and expertise to a wider range of topics in data management. The interview was a fascinating lesson in “innovation through evolution”. 

Finn: Your company SiaSearch started in the automotive sector, more specifically in autonomous driving. Do you have a vision for what the sphere of mobility will look like by, say 2050? 

Clemens: I think it would be a bit presumptuous to just project the future of mobility that far out. But that's also one of the great reasons to work on mobility, it’s an extremely dynamic field - and has been for the last 200 years. Today, I would say three main themes emerge: firstly, mobility is becoming more electrified. That of course has been greatly impacted by the work that Tesla has done, independent of what you think about them. Secondly, sharing is becoming more crucial. This includes thinking about what distances you should cover with what kind of device. And lastly, what we are working on, autonomy. 

And, I think autonomy is  a particularly exciting piece because it enables so many more dimensions. It affects urban planning, individuality, time and space usage. What we can say now though is that people have been a bit optimistic about the speed at which we could arrive at full autonomy. We won't just suddenly have robot taxis and this grand vision that McKinsey pointed out in their reports. Instead it will come through a whole bunch of continuous improvements, solving autonomy in a constraint type of environment. So starting from driving on the highway, which we sometimes already do, but also of course in the unmanned areas of logistics, last mile delivery, in construction and mining. These are all areas where autonomy is partially already in operation.

It makes you very humble because you think that you know a lot about a specific domain and then are forced to scrap another one of your initial assumptions.

Innovation through evolution

Finn: Even before SiaSearch was founded, you went through a couple of company ideas. Would you want to describe the path you took and where you ended up today?

Clemens: Originally, we started - in early 2018 - working with different companies on autonomous driving solutions. That was all very research-intensive, working with companies like Volkswagen and Bosch we ran implementation solutions with deep neural networks on vehicles. Those were basically paid research projects or publicly funded projects, which are partially still running. For us, it was very much the exploratory phase where we tried to figure out where the biggest potential was for a startup company.

Finn: What do you look for at that point?

Clemens:  You're trying to figure out what is an element in the value chain that you can really productize and where you can build something that will solve a specific customer problem. But it also needs to be generalizable and interesting for a large number of customers.

So very early on, we had the idea of building an independent testing company to run validations on behalf of automotive companies. It is a super critical area and automotive companies with an expertise in building cars have not yet the experience in software building and testing. So we started out in that direction but then, pretty soon we started to realize that because it is so critical companies keep testing very close to their core development and are not as open to collaborating with third parties. So these companies liked our ideas and wanted to talk to us - but at some point we decided it would be too hard to go into their development cycle. So, eventually, we discarded that angle and turned towards what we are currently working on.

Finn: But in a sense you remained within the field of testing?

Clemens: Yes, we actually came across this in previous projects and POCs where we noticed that most of the testing with autonomous vehicles’ software is done on actual recording. So companies spend a lot of money recording real world driving data, which they then use one to train their systems. And one of the things companies actually are most interested about is now that they have all of this data lying around, how do you extract things that you really want to use for testing? To make it a bit more graspable, an automotive company has many cars running around and they're interested in very specific aspects of the data such as, situations where a pedestrian is crossing the road or, situations where it's dark or where they have lights against them. And we set up a data management platform that enables customers to structure and search this data for these crucial driving situations.

Finn: And what are you working on right now?

Clemens: In the past couple of months we actually started to go through another evolution which is probably an epitome of building a deep tech company. So, during the Corona crisis car companies went into crisis mode themselves and some projects and collaborations got postponed. So we stepped back and looked at the technology we built. And we started to realize that the problem we solve is actually very present in all kinds of companies that employ machine learning engineers: massive amounts of data and a lot of time and nerves are lost on skimming through this data and structuring it. So, they are all looking for ways to structure the raw data and that is a field that has gained a lot of momentum in the past couple of years. And we've actually been previously approached by companies outside of the automotive industry, like in ocean photography, mining and surveillance, for example.

Frustration tolerance

Finn: Going through these evolutions, as you just called them: hypothesizing a product or service idea, trying to validate it, noticing that this is not yet the value proposition you want to run with. Is that one of the benefits of being a founder or is that mostly exhausting?

Clemens: Both. It’s very much an unavoidable part of founding. Very few companies start with an idea and just execute it. What you are looking for is this ominous product-market-fit. So trying to really find a solution that a company is trying to solve and that you are then able to scale. It's probably also the most interesting aspect of starting a company. It makes you very humble because you think that you know a lot about a specific domain and then you are forced to scrap another one of your initial assumptions. So it keeps you in this mode of being very curious, very open.

But it is certainly also frustrating as you said. You need to be confident and have this conviction about the vision you are trying to sell. But at the same time you need to be ready to adjust. And balancing this out is tough.

Finn: Surely, that must be a very hard line to walk. Both having conviction but at the same time being open to change your mind?

Clemens: It’s a process. I mean one of the things I realized as a founder is that you need a high degree of frustration tolerance, you will definitely face some walls, you will have setbacks. And that is one of the things I learned from one of the managers in my first corporate job, and at the time I didn't really understand it that well: if you have a high tolerance for ambiguity that really helps a lot in a startup environment. If you have a desire for everything being planned and you being able to write a detailed plan and then it just becomes reality, then you will definitely be disappointed. No business plan survives first customer contact.

How to identify great entrepreneurs-to-be

Finn: You have a cofounder, SiaSearch’s CTO Mark. How did he get involved?

Clemens: After we had encountered our initial product idea, we wanted to build a prototype and try this idea out with initial customers. But I came very much from a generalist angle and so I needed someone with solid credentials in building tech teams. And then we convinced Mark to join from his PhD in robotics. But I have to stress that his expertise is really only the necessary condition. When looking for a cofounder the personal match is absolutely crucial. You need a similar set of values and a similar vision for the product you are building.

Finn: How do you identify a great entrepreneur-to-be?

Clemens: Curiosity, openness, having a strong sense of direction, resilience, wanting to lead people and wanting to take responsibility. All of that is important. And then I’d say you need to be able to make decisions under uncertainty. That’s actually one of my favorite interview topics because I think it’s something that people typically say in an interview that they thrive on uncertainty or under stress. But I think not a whole lot of people really do. When actually confronted with a situation like that, people look for guidance and are not extremely excited about the situation. 

Finn: That was really interesting, Clemens. Thank you so much for joining me, I really appreciate it. Best of luck with SiaSearch!

Download pdf