Yuval Boger interviews Jannes Stubbeman, co-founder and CEO of Aqora, a platform aimed at creating a “Kaggle for Quantum” by providing an operating system for both in-person and online hackathons in the quantum computing space. Jannes and Yuval discuss the origins of Aqora, how it supports hardware vendors in organizing impactful hackathons, the intricacies of benchmarking quantum use cases, the challenges and logistics of hosting competitions, and much more.
Transcript
Yuval Boger: Hello, Jannes, and thank you for joining me today.
Jannes Stubbeman: Yeah. Hi, Yuval, thank you for having me.
Yuval: Of course. So, who are you and what do you do?
Jannes: I’m Jannes, co-founder and CEO of Aqora. And with Aqora, we want to build something like a Kaggle for quantum. So being an operating system for in-person hackathons and also enabling online competitions, which, in the end, should be an easy way for companies to explore this technology by crowdsourcing these solutions. And maybe a bit more. I’m German. Right now, you’ll find me here in Berlin. We have set up Aqora in Paris, so I’m somehow living between these two cities these days.
My background is actually in computer science, which I studied mostly with a focus on machine learning and then quantum computing.
I got exposed to the topic in my master’s thesis actually when I was simulating the Google supremacy experiments with classical neural networks, these Boltzmann machines.
And yeah, since then, I’ve been staying active in this field as president of PushQuantum, which is a student’s initiative at the Technical University of Munich, but I have worked directly in the field only since last year, since we kicked it off. Actually in November, so it’s almost exactly one year ago that we started with the Aqra.
Yuval: Wonderful. Let’s assume I am a hardware vendor and I want to organize a hackathon. What do I do?
Jannes: It depends on what you can and want to bring to the table yourself.
Actually, our background, the origin of Aqora is really in-person hackathons that we were running as QuantX, mostly my co-founders, Alexandre and Elvira, who had co-founded QuantX and been running some of these hackathons over the past few years.
And so we have expertise running them in-person or virtually or remote.
I guess it depends on what you want to achieve also in the end.
I guess as the vendor, you want to expose your technology, right?
Yuval: Yes. Let’s assume that I want to expose my technology. Let’s assume that I have a challenge or a couple of challenges that I think are representative of the good things that my technology can do. But let’s also assume that I have no experience and I have an audience. I know where the students are or where the participants would be, but let’s assume that I have no experience in running a hackathon. What can you do for me? How does it work?
Jannes:Aqora as the platform provides you, let’s start from there, is basically this interface where you can provide your use case in an accessible way. And we put a lot of emphasis on these benchmarked use cases because there’s often a question of how to evaluate these solutions. So we will provide you with this. We would support you in bringing these use cases online.
And then, it’s the question whether you want to have this as a public event virtual to the global audience of users on our platform, or whether you want to have a closed event also could be virtual or in-person.
If you decide for the in-person, we would need to find the venue and the catering and everything. This is a bit of the general event organization stuff that we have expertise within and that we can help with, though it’s a bit more complex to do it outside of Paris for us, but still we could do that.
We focus on these online competitions, providing these use cases on our platform, making it available to your audience that you could either invite or make it public.
In addition to our website where your use case is displayed, we also have our command line interface to check out this use case and set up your virtual environment because what we have experienced from in-person hackathons in the past, everyone has their version of, well, say Qiskit or any other Python SDKs.
With us, you could also provide a preset environment, maybe having an example of your technology and then people could take it from these template solutions and start working on their submissions and then submit them.
You can see how they would score and maybe have other measurements to decide then who would be the winner.
And yeah, and again, it depends on how you would have this format in the end as the hackathon really.
You can have also the ceremony and everything.
But I guess what we can mostly help you with with Aqora, as I was saying, is really bringing the use case online together with some initial template so that it is available and accessible to this global audience of quantum experts who have an easy method of working on it, evaluating their submissions and also sharing with each other their environments easily.
Yuval: I’m curious about hackathon participants. Do you see professional participants, meaning participants that go from one hackathon to the other and that’s sort of their hobby, just do one after the other?
Jannes: Oh yeah, actually, yes, absolutely. I guess we have seen this with in-person hackathons, especially if we have them in the same city – there’s the annual one in Paris with QuantX, but also now with the virtual competitions.
Since we have made this transition this year, we had our first competitions organized end to end and we’ve seen that the same participants show up again and also with the ones to provide them very good solutions.
And it’s one aspect why we actually do Aqora, like one of them, because what we wanted to do with Aqora is also that the profile should be something like a reference profile.
It’s not just about exploring this technology, but also for the experts that they can demonstrate, that they can transfer their theoretical knowledge to real world applications, well, on the small scale these days, right?
But that they can come to one hackathon or competition after the other and by that build up their portfolio and later be able to demonstrate which kind of use cases they’ve been working on and also maybe how they rank in this.
Yuval: How many hackathons have you performed in Aqora?
Jannes: So as Aqora in-person hackathons, it was one so far. This was our initial test of the platform in May actually in Paris. As I was saying before with QuantX, this was more my co-founders, I think they organized, if I’m not wrong, was it eight, I think that have been performed over these past four, five years.
And now we are running one big online challenge, which is the QInnovision World Challenge 2025, which has multiple use cases that we are running since September and until the beginning of December.
We have been running two competitions end to end, like virtual ones, one with Ingenii on clinical trial optimization and one with NovaceneAI on malicious login detection.
And now we are running a next one that we just kicked off last week with Quantum Signals on predicting stock price trends.
And we will also actually run a hackathon in-person at the Q2B in San Francisco together with IBM.
Yuval: In your experience, both in Aqora and before that, what makes a good hackathon? What are the keys? Again, I wanted to run a hackathon or to organize a hackathon through you. What do I need to do to make it really impactful?
Jannes: We can decide between in-person and virtual again. With Aqora, we are focusing more on the virtual ones, but we also see it as the operating system for the in-person ones. So I think in-person, well, it’s a lot about the logistics so that everyone really has a good experience. It’s also about matching the right teammates before so that there are balanced teams.
And then actually, I think one of the important parts really is also to have this closing reception set up in a good way, because that’s where a lot of networking happens after the event. And it’s really something that I experienced in these in-person hackathons that I participated and co-organized is at the end of the hackathon, there’s a lot of energy and a lot of enthusiasm. And I think it’s crucial to somehow not let this die off. So this was also one of the reasons why we started Aqora, that there can be this in-person hackathon to create a lot of enthusiasm, but also enable participants to continue working on their solution afterwards.
What we see for virtual competitions, what I think is crucial, is more crucial than in the in-person setting, from what I’ve seen so far in this last year, is to have these kind of metrics. So that we have a real strict benchmark on your use case, because then people are really guided towards something they have to optimize, so they have their objective to optimize. I think if it’s more open, it can also work and people love to be creative, but I think it’s a bit more difficult to provide the structure without this in the virtual setup.
Yuval: You mentioned earlier that one key is matching the right teammates. Does your platform help in this matching process in any way?
Jannes: Yeah, that’s a good question. So we have the plan to have easy features to ping others and ask for team up. So far there’s no explicit feature to do these asking others, at least. We have this concept of organizations, so you can team up, you can hand in as an organization, which can be your peer group. And well, we have discussion forums, so you can reach out to others and ask to team up. So well, to some degree, yes, but there’s more work that we want to do for making it even smoother.
Yuval: How are quantum hackathons different than classical hackathons? I mean, the concept of a hackathon existed before quantum computing. Why is there a need for a quantum specific platform?
Jannes: Oh yeah, yeah, yeah. That’s also a very good question that we asked ourselves before we started this, because as I was saying in the beginning, we see Aqora, at least in this state, there’re other plans for the future, but we see it as something as the Kaggle for quantum. And many people know Kaggle because it’s this very popular machine learning data science community platform. And actually when we started Aqora, we were wondering why is there Kaggle, but also why is there a Hugging Face? So we wondered why Kaggle did not manage to become something like a Hugging Face. And there might be different reasons, but we also met the Kaggle co-founders.
And one thing we understood is that with these online communities and platform, it’s often about one core value that they can share. And I think with Kaggle, it’s nowadays, I think on their website, they also present themselves as the data science platform. So I think this already gives the hint that it’s a lot about digesting the data and then probably using more ready-made algorithms while Hugging Face, for instance, is focused on building these LLMs. So you need more expertise on the specific algorithms.
And this is then also what we see with quantum that we need something where really the quantum computing community can gather and work on the level and with the state of the current technology.
Because this is then also the crucial part that of course, at the moment, we, for instance, when I was talking about benchmarked use cases, so what you can see on our platform, if you provide such a use case and people hand in their submissions, you can see a single score, which can already give you a good direction of what you want to do. And if you have set up correctly, that’s something that will certainly help you a lot to really understand the solution.
But I think the more interesting part in this phase of quantum computing, and we want to go there is to show some kind of more graphs and also show multiple metrics. It can be your business metric and you might want to see how does my score, I don’t know, depend on the size of the problem instance. If you do portfolio optimization, you might not be able to optimize a portfolio with 1000 or more assets these days. Maybe you can do one, five, 10, 20 assets, and then see how your score, which could be, I don’t know, runtime optimal value, the business metric you want to optimize all changes.
And at the same time, if you can then overlay how, I don’t know, some quantum resources you would need to develop that, like say X qubits for five, 10, 20 assets in my portfolio, then you can make these kinds of projections.
Because then you can see, okay, well, given the technology roadmap, and I can see this that in a few years there might be so and so many qubits or not with this and that connectivity, you could start making your predictions.
And I think this is where then the power of our platform really can lie in that you can explore this technology at low costs today, not to expect the production ready, high impact solution for your business today, but to understand how you can have your roadmap prepared.
And you can do this on your data and your use case and very much tailored.
And I think this is something that other platforms do not have the focus on, but that we have our focus on.
Yuval: In your experience, do most hackathons run on real hardware or do they stick to simulators?
Jannes: Yeah, so far they run on simulators.
The biggest so far we had, I think was one, a solution that was run on, I don’t know if I say the right name, but NVIDIA was once joining as a provider in Chicago last year and they made, I think something like 40-ish qubits available to the team so they could actually run some algorithms with these many qubits. It was pretty cool.
And actually, so with the hackathon that we will host with IBM at the Q2B in San Francisco, IBM has plans also to make the real hardware available to the teams, not in the first stage, but in the second stage.
And yeah, so we can see sometimes there’s some access to real hardware or at least a high number of qubits, but not always, yet.
Yuval: Does the use of the platform start before the hackathon? So for instance, do people use the platform to pre-install the environment or maybe their test examples just to see that they are at the right level and that they are successful in running something?
Jannes: Yeah, so we provide actually one tutorial where people can play with to calculate the ground state energy of the H2 molecule, which we always suggest to test once.
And yeah, I mean, another thing that we learned from running these in-person hackathons, you should make sure that people have all the accounts and access to the platforms that you want to use at the hackathon before the event. Otherwise, you waste two or three hours just walking around and making sure everyone has access to everything. So yeah, we provide early signups with the Aqora account and then everyone will be prepared at the event.
So I guess the real hackathon, or if you look behind the scenes of the organization, it starts two or three months before so that everyone can be ready set up at the event when it starts.
Yuval: You’ve been doing this for about a year in Aqora, if I understand correctly.
Jannes: Yes.
Yuval: What do you know today about hackathons that you didn’t know 12 months ago?
Jannes: Oh, well, what I didn’t know about hackathons. So for the in-person hackathon that we were running with QuantX, well, I’ve seen bringing all these different parties together. So the format of QuantX might be a bit special, but I think in a good way is really that there are participants like the quantum experts working in teams on a use case to which they’re matched. And there’s also one dedicated hardware provider matched to them. And this requires a lot of upfront logistics. They are making sure that everyone will agree to be there and will be okay also to be there matched to the other teammates and hardware and use case providers.
So this is something that I could learn that this is a lot of overhead. And I think for the online competitions and our Aqora with the focus on this platform, it’s a lot about in these terms about solving this chicken and egg problem. Should we first have a competition online or should we first have something where there’s enough quantum experts so that they can start working on something? But yeah, so I would say it’s really about these logistics that require a good amount of work, consistency, and ideally a good network. So that we have good co-founders that are very well connected and have done this now for a few years because this certainly makes a few things much easier.
Yuval: Wonderful. Jannes, thank you so much for joining me today.
Jannes: Thank you, Yuval, for having me.