Transcript | Helping Quantum Computing Startups Code— with qBraid Listen If you build an easy-to-use portal to the world of quantum coding, eventually, companies are going to take notice and want to offer the environment to their teams. Learn how you can interact with quantum computing hardware and software of all types, either on your own or with your teammates to solve real business problems, using qBraid. Join Host Konstantinos Karagiannis for a chat about this evolved tool with Kanav Setia.Guest: Kanav Setia, qBraid Listen Topics Digital Transformation Konstantinos Karagiannis:If you build an easy-to-use portal to the world of quantum coding, eventually, companies are going to take notice and want to offer the environment to their teams. Learn how you can interact with quantum computing hardware and software of all types, either on your own or with your teammates, in this episode of The Post-Quantum World. I’m your host, Konstantinos Karagiannis. I lead Quantum Computing Services at Protiviti, where we’re helping companies prepare for the benefits and threats of this exploding field. I hope you’ll join each episode as we explore the technology and business impacts of this post-quantum era.Our guest today is the CEO of qBraid, and a repeat guest, Kanav Setia. Welcome back to the show. Kanav Setia:Thank you, Konstantinos. Pleasure being back. Konstantinos Karagiannis:A lot has happened since you were on over 40 episodes ago. As we’ll get to, I do end up seeing you though, so that’ll come up a few times. But for those who didn’t hear that episode — and it was a good one — let’s give an intro to what qBraid is for listeners. Kanav Setia:QBraid is a one-stop platform where you can access all the different quantum software and hardware from many companies around the world. If you want to access Qiskit, Cirq, AWS’s Braket, Rigetti’s pyQuil, Xanadu’s PennyLane, you name it, we have that software available. We also support more than five quantum computers available on our Lab platform. Essentially, you just come in, everything’s pre–set up, you open a notebook and start finding the code, start submitting the quantum circuits, and that’s basically what we do. Konstantinos Karagiannis:People should be familiar with this concept. If you’ve ever done a tutorial using notebooks, you could jump in and actually do things in code. It’s not that dissimilar from Google’s Colab — you could just go right in there and do something inside of a window. I’d love to dig into some reasons for why you’re back on. For starters, this isn’t a tool for just practicing coding, although it’s great at that. Real businesses are relying on this. Can you give our listeners a sense of some examples of how real businesses are launching products on this platform? Kanav Setia:As you mentioned, qBraid is similar to Google Colab, as in, it has a field similar to that of Jupyter Lab and the Jupyter Hub platform. In fact, we leverage the Jupyter infrastructure and build on it and make it super friendly for quantum-specific tasks. We build custom extensions that fit right into the Jupyter Lab platform that provide you access to Python Environment Manager.We built our custom UI for Python Environment Manager that allows you to segment your software into various Python environments. And we have many custom Python environments that we shift by default, so people can just open a Python environment that is custom-built for AWS services, and they will get all the software from AWS — similarly with IBM and Google, Microsoft, Intel and UniQ. Now people are starting to build their software on qBraid and sometimes also launch it to their potential customers, do the demos and so on.We’ve been working with many quantum companies to launch their custom Python environment so that their tools will be readily accessible to qBraid users and their end customers. We work with these companies and build a custom Python environment with their software. Any user can just log into qBraid, install that Python environment and get access to all the quantum software from a particular company and open their notebooks, and can readily execute the circuit. Further, within these Python environments, we make sure that these are configured to connect to various quantum hardware.There’s no separate making accounts required for AWS or many other companies. You simply come to qBraid, and we have our own credit system where you spend money to buy these credits. Once you buy these credits, you can turn on Quantum Jobs, which is basically one line of code or one button in our UI. You press that button, and suddenly, all the code that submits a job to quantum computers starts executing and your job management happens automatically. Essentially, as an end user, if I want to get access to a particular quantum company’s software, I can get their Python environment, turn to Quantum Jobs and I will be set. I will open their tutorial notebook and execute it. Because all the software is present, all the code will be executed. And if there are any jobs to be submitted, they will go through to the quantum computer where I want to send those jobs, and I’ll be able to get my results back. Konstantinos Karagiannis:So listeners can visualise what that looks like, you have the credit system, which I notice is different from last time, definitely. And that makes it simple, because like you said, you don’t have to have API keys set up for running different jobs and paying like that. Let’s say you buy a bunch of credits and you pick one of these preselected notebooks from a customer. Would that be something where they’re charging you a split? Some of the credits go toward running the machine, and some of the credits go toward them as “You’re running our software”? Is it a whole ecosystem and economy like that? Kanav Setia:We are starting to explore that, but the field is still early where our focus is mostly on learning and building the right thing. Right now, as you may already know, many of the listeners might already know, quantum computers are not where they need to be to offer real value to the end customer. There isn’t any software in the world that you can build that can run on these quantum computers and provide you meaningful advantage over classical software and hardware. There’s not a lot of value in building an infrastructure that allows anyone to build an application and launch it for their customers so that their customers can launch it for the API. So most companies are just open sourcing their software and trying to build community around those. Konstantinos Karagiannis:I wanted to make it clear with the credits and everything— where those go ultimately. They’re going to running these not-fault-tolerant quantum computers in the background. Kanav Setia:Mostly, there isn’t any split or revenue sharing with quantum companies. But this is something that is being actively discussed with various of our partners. In the future, as the business grows, yes, that’s certainly something we would do where if a lot of people are using a particular quantum software from a given quantum company, we will provide revenue sharing with them. That’s a model we have started to think about. Konstantinos Karagiannis:For those who want to implement qBraid in their organisation, I’d love to talk about some of the features — some of them are new since last time you were on — that help companies do that internally for managing jobs and functionality. Can you talk about what that looks like if you have a team and you want to implement this internally. Kanav Setia:We recently shipped another awesome feature, which basically gives organisations an ability to make an organisation page on qBraid, and you can buy credits in bulk. Let’s say you spend $25,000 and you have a whole lot of credits. You can have users come in and join your organisation. You can give them roles based on different permissions, and then you can transfer credits as they go. Automatically, this problem of overspending gets sorted out because you can transfer credits $10 at a time or even $1 at a time. If the job requires more credits, then it won’t go through unless you top it up.That problem is resolved. Any organisation, they make their page, give their developers access to quantum computers. Further, we’re starting to build features where they’ll be able to manage the classical compute for their developers. Let’s say you are running some huge tasks and you want to simulate them before running on one computer. We are in the process of giving people access to a scalable compute. We already provide three, four options: You can launch a four-virtual-CPU machine, ten-virtual-CPU machine, another GPU machine, but we’re planning to grow this to maybe 20 or 50 options where you’ll be able to have access to Nvidia’s A100 or B100. And as you go, you can just start up any of these machines, run simulations, test your results, turn off those machines, and you will be billed through qBraid credit — a single credit system for all your QPU, CPU and GPU needs. Konstantinos Karagiannis:And what about privacy and security? I’m assuming if an organisation creates one of these, they definitely don’t want to be sharing it. They don’t want this in some community setting. How do you guarantee that? Kanav Setia:We generally follow very good practices for all our data. We make sure that it’s segmented from any other user’s data. Beyond that, if there’s a requirement from an enterprise that they further want their data to be encrypted, that’s certainly something we can work with organisations about and provide them this service. It’s general security practice that your data is segmented and sandboxed and doesn’t interact with any other common data. Konstantinos Karagiannis:Definitely, when you get into the realm of developing IP, these things start to come up. Kanav Setia:This is going to be one of the most important things as the field develops — the issue around IP. Another important thing is, as the scale of usage in an organisation grows, we’re also happy deploying our software infrastructure to their private cloud, so that way, the data never leaves their private cloud. We would basically manage the software infrastructure for them.We’ve done that with a couple of organisations where we white-labeled that product and deployed it on AWS because they didn’t have a private cloud. But essentially, the mechanism is very simple: We white-label a product and deploy it on your AWS clusters or your private cloud clusters. Konstantinos Karagiannis:With these private companies doing that, do they then get a different level of support if they have problems executing something or errors and they need technical help? Kanav Setia:Based on the contract size, we definitely bundle in some technical support for a certain number of hours. It depends on from contract to contract, as in, what are their requirements, how much support they’re going to need. We certainly provide that service as well. Konstantinos Karagiannis:When you were last on, this sounded like a great way to learn and explore, and now it sounds like a real tool ready to hit the road — actually put some quantum miles on. It’s nice to see that it’s growing. And it looks like you’re having an impact in academia too. Can you talk about QuSTEAM? Kanav Setia:QuSTEAM is one of our biggest customers, and they are the ones who white-labeled our product and have deployed it for their own organisation. They’re planning on being at 300 universities in the next five years. What they’re going after is this need of developing content for training future quantum physicists and scientists. They have an incredible team of 50 to 70 professors at more than 25 universities, and they’re churning out content.What we are solving for them is, essentially, you have this content. How do you deploy this content in a scalable fashion? We make sure that they have the right tools. We’re building some custom applications for them that will allow professors to build courses within qBraid Lab and deploy it for their users so they can publish the courses and then they can add students to certain courses, and all that content appears in students’ qBraid accounts, and they can open it and start running their code, start doing exercises and so on. Konstantinos Karagiannis:Does this look similar to what the blogs look like, where it’s just a mixture of instruction and code? Kanav Setia:It’s basically, you can build notebooks in qBraid Lab and bundle 10 of them, and you can deploy it on the blogs section. Just imagine bundling up ten blogs together, and now you have a course, and you have an outline and it’s the same platform, but now you can just bundle up blogs and you can deploy it as a course. Konstantinos Karagiannis:Then they have a place to put their actual coursework into an interactive mode. Teachers can see, is their code running — all that stuff. They’re taking the white label you gave. Are they building any new functionality in the back end, like some teacher dashboard with scoring and things like that? Kanav Setia:Essentially, the technology part is what we are taking care of. It’s making sure that if they have any custom needs. The two products might diverge because the needs of the industry might be different from what is required at academia, which is why we thought this might be the best strategy. While they are selling the content, they can also sell it along with qBraid access. That was the whole point of having a white-label solution where now they can also have a course builder, and we can deploy that. But there’s perhaps not a huge need for a course builder for everyone on the platform. Konstantinos Karagiannis:And now you can take this bragging point and bring it back to companies. You could say, “They’re using this now to teach in schools. You can use it to teach your employees” ultimately, in some way. Kanav Setia:One more cool point is, we can also use content and deploy it in the industries as well because now this content has been validated, and many companies, as you may know, they’re just starting out. Their first step looks like, “We hear so much about this quantum field. We want to know what it is all about. Can you teach our engineers?” We can literally take similar courses and use it for training. And of course, we could need QuSTEAM’s permission. Konstantinos Karagiannis:I’m sure they’re doing some of the IP there, but this is all one place, and it works for everyone. One thing to keep in mind is that this is such a challenging field for people to learn. It’s great to see something all bundled up and easy. The other thing we talked about last time is developing these environments for use. It’s no small feat. It’s the reason people started using containers, because it was a nightmare to get just the right version of each thing pip installed, and then good luck. What kind of nightmare comes after pip install?It’s impressive. That’s why I wanted to have you back on and see how this has evolved for business use too.We’ve been spending a lot of time together lately meeting for the U.S. National Science Foundation competition. For those who don’t know, this is all public information, but the University of Chicago, our Quantum Crossroads team, was announced as one of the 16 finalists in this first regional innovation competition. They span all types of technologies, and we’re one of the finalists as quantum for things highlighted in the CHIPS and Science Act. What does being part of a big multiyear competition and possibly winning something like this mean for a company like yours? Kanav Setia:It definitely, first thing, brings us so much credibility. It’s such an amazing privilege to be part of an incredible team led by CQE and UChicago. It shows that we must be doing something right that we’re getting attention from such amazing folks. That’s one big point. The second thing is, we get to be part of an incredible team that’s taking up one of the biggest challenges in the field. Not just one — they’re going in so many different directions. They’ve got three cores: There’s R&D, innovation and workforce. All of these are going to be some of the biggest challenges that would need to be overcome to make the quantum dream a reality.To be able to work on these amazing problems, and being part of the solution and struggle with this and make things happen, that would be the most fun thing. If the grant comes through, we’re looking forward to contributing toward workforce development, which is what qBraid is leading, and contributing in other places as well. Konstantinos Karagiannis:I’m involved in more of the R&D track, but we are all like one big team anyway. What impresses me is that whatever we accomplish — assuming we win this — in the Chicago area, it would be a great model for what can be done all around the world to make sure that quantum isn’t limited to just certain folks in certain areas, all sorts of issues like equity and things. It’s an exciting project to be part of. On that note, recently, you won some other grants. Can you talk about those? Kanav Setia:We recently were awarded two main grants. One was Q4Bio from the Wellcome Leap organisation, a grant focused on the use of quantum computers for biological problems. It’s up to $4 million, structured in three phases: The first phase is up to $1.5 million. The second phase is up to $750,00 and the third is up to $2 million.We submitted a proposal. We have Professor Laura Gagliardi from UChicago, Professor Troy Van Voorhis from MIT, research scientist Yuri Alexeev from the Argonne National Laboratory and Professor Ray Samuel from the Medical Clinic of North Texas. We’ve got a solid team, and the project idea is quite fascinating.We came together as a research team and discussed all the smaller pieces we had in our individual results. It was amazing to see how they fit in nicely together. You have these bigger protein molecules, and what we proposed was simulating an entire protein molecule. Even on the biggest quantum computers, even in 30 years, it might not be feasible because the protein molecules are so big. You can segment these using crude fragmentation techniques, which are available in the classical world as well. You have a bigger protein molecule, and you use crude techniques to chop them up, and you end up with these smaller segments.It turns out these smaller segments still may be big for near-term quantum computers. They might be doable on future quantum computers, but for the near term, they still might be big. And that’s where we have techniques from Chicago, Professor Gagliardi’s group, and Van Voorhis’s group, where they have further quantum segmentation techniques. You take these smaller segments, fragment them further and then simulate them on quantum computers. The smaller fragment is a fermionic system. And the normal techniques you use to map a fermionic system to qubits is something called the Jordan-Wigner transformation.During my Ph.D., I worked on an encoding called superfast encoding. When you use that encoding, you have to use more qubits, but that provides you more efficiency. You end up requiring lower gate counts and a lower circuit depth. But on top of that, you end up getting free error-correction capabilities. We ended up fitting the entire pipeline together, and the reviewers must have liked it, which is why we ended up getting the grant. Konstantinos Karagiannis:That’s pretty impressive. I was hoping you would talk about the use case. That’s cool. Protein comes up all the time. If you hear quantum and bio, it’s always going to be protein. One of these days, there’s going to be a quantum protein bar that we’re all going to be eating because it just keeps coming up all the time. Kanav Setia:A particular use case we’re going after is these proteins that end up in markers for Alzheimer’s disease and Parkinson’s disease. There are these proteins called beta amyloid, for Alzheimer’s, and alpha synuclein, for Parkinson’s, which apparently react with metals and form these flakes, which are known to be markers for Alzheimer’s and Parkinson’s. And the way this interaction happens and the way they get formed, it’s still relatively understudied and not very well understood. This is what we are going after. Konstantinos Karagiannis:That’s good. I realised it wasn’t protein you eat — never miss an opportunity for a joke when you can take it. As far as the error correction, did you want to say something about that? That’s an interesting approach. Kanav Setia:The error correction, the original idea was that, if you have a fermionic system and you’re using Jordan-Wigner, you have this order-in overhead. And if you were to use more qubits, you could map qubits in certain ways that it ends up causing only order-constant overhead. You reduce order-in to order-constant, but you are using more qubits. That’s essentially the idea from my Ph.D. that I observed this, that you were using more qubits. If you’re using more qubits than the number of fermionic modes, essentially, you’re using a subspace.Apparently, even the formalism for using this subspace out of the entire Hilbert space was very similar to the stabiliser formalism used in error correction. It was a big conjecture that I had: If you’re using the same stabiliser formalism, you can use the same formalism to do error correction. I ended up proving that. Then I worked with Sergey Bravyi, whose original idea was this encoding where they brought down the cost. We ended up working together, and we proposed a new encoding where we showed that if certain conditions are met, you can always guarantee single-qubit error correction. Konstantinos Karagiannis:There’s no such thing as wasted effort. In this case, everything makes a return appearance. And then you have something going on with NSF ExpandQISE. Kanav Setia:That’s another grant that we received in collaboration with North Carolina Agricultural and Technical State University. We run a decent number of programmes with them. We’ve had two courses with their undergrad students. We ran a workshop for their training, a workshop in quantum computing for their faculty. We’ve had a wonderful experience. We ran the workshop, and we trained the faculty. And once they were trained, we ended up applying for a quantum grant together with the faculty. And that’s the ExpandQISE grant from NSF, which is for $1 million per year for five years. And that grant is going to focus on two main themes: One is quantum chemistry. Another is aerospace engineering. Konstantinos Karagiannis:That’s great. You can see that Kanav and his team don’t just sit around waiting for users to sign up. They go out there and make some opportunities happen. This is impressive stuff. I urge everyone to go to qBraid.com, link in the show notes and open an account, and see what it’s like to play in that environment. And if you’re part of a bigger company with a team, as Kanav said, there are ways to make it something that’s useful internally. It’s definitely worth checking out. Thanks again, sir. I’ll see you in one of our very many meetings after this. Kanav Setia:I believe there’s a meeting tomorrow as well. Konstantinos Karagiannis:Absolutely. See you there. Kanav Setia:Thank you so much for the invite. It’s a pleasure being here. Konstantinos Karagiannis:Now it’s time for Coherence, the quantum executive summary, where I take a moment to highlight some of the business impacts we discussed today. In case things got too nerdy at times. Let’s recap.QBraid lets you access quantum computing software and hardware from most major providers all in one interface. Similar to when using other coding notebook models, you just load an environment and start typing or modifying code. No worrying about installing dependencies and the like. The site now features qBraid credits you can buy too so you don’t have to worry about having API keys to access real quantum hardware on the back end.While the site began as a place for anyone to learn and experiment, it’s now evolved to also support more enterprise-level use. You can buy credits in bulk for your team and then divvy them up by role, project and so on. You can even impose limits on jobs to prevent accidentally running up a huge bill on multiple shots, for example. There is segmentation in place for these types of corporate users, and qBraid even supports installing in private clouds if needed to meet some requirements. Corporate users enjoy bundled hours of technical support too.QBraid is also supporting the world of academia now via QuSTEAM. The platform will be available to 300 institutions over the next five years, allowing professors to create complex course curriculums by adding modules. These modules can be mixtures of text explanations and executable code windows, which should be familiar to anyone who’s used Jupyter Lab or Google Collab tutorials, for instance.The team at qBraid is still involved in research too, recently winning a grant for a use case featuring the simulation of proteins by chopping the problem into sizes manageable for today’s quantum systems.If you’d like to try qBraid either on your own or through a corporate account, check out the link in the show notes.That does it for this episode. Thanks to Kanav Setia for joining to discuss qBraid, and thank you for listening. If you enjoyed the show, please subscribe to Protiviti’s The Post-Quantum World, and leave a review to help others find us. Be sure to follow me on all socials @KonstantHacker. You’ll find links there to what we’re doing in Quantum Computing Services at Protiviti. You can also DM me questions or suggestions for what you’d like to hear on the show. For more information on our quantum services, check out Protiviti.com, or follow Protiviti Tech on Twitter and LinkedIn. Until next time, be kind, and stay quantum-curious.