Transcript | Coding Quantum Computers for Free

It’s amazing how many quantum programming languages and interfaces exist already. What if there was a way to learn and use all of them without installing anything? qBraid is making that possible with a site you can join for free today. Find out what to expect on your coding journey 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. I’d like to welcome Kanav Setia. qBraid is one of the companies that was involved in Duality, which was the first accelerator for quantum companies. There were six companies chosen, and his was one of them, so we’re going to talk about that and what they’re doing. Welcome, Kanav.

Thank you for having me. I’m super excited to be here.

If you want to give us a quick taste of what being part of an accelerator did for your company? I was curious about what kind of boost that gave to an up-and-coming quantum company.

It has been super helpful. The kind of connections that we got, the access to the network, all of those things have helped us quite a lot. Especially the access to the network — this is something that we would have never imagined that we would have access to. The accelerator itself is a joint programme from the University of Chicago, UIUC, Fermilab, the Argonne National Lab and the Polsky Center for Entrepreneurship and Innovation at the University of Chicago.

When they do programmes or they put together events, they get to invite people from all these organisations. They also get to claim that this is the U.S.’s first startup accelerator focused on just quantum, which means any VC who’s interested investing in quantum will check it out, and any programme where they want access to quantum startups, they will reach out. I think that the support we’ve gotten from Chuck, Preeti and many others at Duality Accelerator, that has been unparalleled. We’re very fortunate and thankful for all the support we’ve received.

Yes. It’s exciting. It’s another way that this all feels more real now that you have these kinds of programmes that are helping companies get going, and we do a lot with Chicago Quantum Exchange. I talk to so many people from Chicago, everyone assumes I’m from Chicago, but no, Manhattan. It seems like everyone’s in the Midwest.

Let’s talk about qBraid. You’re doing something that’s pretty welcome: It’s this whole idea of making it easier for developers to get involved and get started. If you want to explain really lightly what the platform’s like, and how it compares to the other ones out there, like Strangeworks and things like that?

The whole mission of qBraid is to accelerate the pace of quantum revolution, and we want to make it super easy to get started in quantum computing, which means you can log in and you can get access to the state-of-the-art quantum software as well as hardware without having to spend a couple of weeks researching into what is required for getting started, and then trying to install a bunch of stuff on your local computer. You just log in. We have software set up for you — we take you from not knowing anything about quantum to running your first programme in under two minutes and under five clicks — that’s one of our taglines.

It’s based on JupyterLab notebooks. We’ve tried super hard to make it hacker-friendly. Compared to Strangeworks, I think that’s one thing I think we excel at — making it super hacker-friendly. This is something that we’ve been told, not something that we’ve been claiming. People have come back who’ve tried out Strangeworks and tried out qBraid, they did mention that they find qBraid to be much more hacker-friendly, as in, you can have access to a terminal, you can create your own Python environment and you can have access to multiple Python environments.

Strangeworks does have a focus on bringing businesses together too. This is their thing, but with qBraid, you go and you create an account. Right now, it’s still possible to create accounts for free and start playing with things immediately. Of course, just to clarify, when you say “hacker-friendly,” you mean in that old-school developer approach. Someone listening might be, like, “What? We can get attacked while we’re here?” No. Not that kind of hacker. This is more like for those who want to hack at the code and get a programme.

Thank you for clarifying that.

We want to keep that clear for everyone at home. When you get there, you have two basic platforms, I see: There’s the Learn platform, and then there’s the Lab platform. The Learn platform, you have courses, and there are more courses coming, correct?

That is true.

You log in, and it lets you start learning in one place, and this is inspired by Jupyter. How interactive is it? What can you do in these courses?

We wanted to make it even easier compared to Jupyter notebooks. For some people, they just want to learn, read and have videos in there, so Jupyter itself is a great platform. We work with people to make courses within Jupyter notebooks. You can have YouTube videos embedded in them, you can have interactive widgets in them and then you get to publish it as a course. This is something you can open in your browser. It’s a single HTML page with the code snippets embedded in them. You click Play to run the code snippets. You can modify them and click Play again, and you will be able to play with it. This is not something that you can save your progress in when you’re coding, but this is something that you get to read, understand the fundamental concepts, understand the code that goes with it and better your understanding through playing with the code snippets.

Who’s developing these courses?

The first one was developed in-house, but the later ones, we worked with a few professors, along with a couple of interns. We’ve taken the material from the courses that were already being taught at universities, and then we’ve hired a few interns to add in code snippets so that it reads with a nice flow.

If you look at the landscape today, there are still very few courses in quantum software. All the quantum computing courses right now being taught at universities, they still teach a lot of the fundamentals and from the math perspective. There’s less coding involved, and this is where we come in. We hire a few interns, and we tell them, “Can you take this course and add in code snippets so people know how this concept could be implemented using, let’s say, Qiskit, AWS Braket or Google Cirq?”

There are some good visualisations there — things you could drag and move to actually see interactively. For some visual learners, especially, it’s the right way to go. Someone who logs in, gets a little experience, plays around with that, then it’s time for them to try and create something, right? That’s when they go into the real meat where this is going to start to be a useful day-to-day environment. They go to the section called Lab. That seems to also have tutorials in there, but it looks like you can create your own code to continue with a problem-solving approach. You want to talk about what you have in there? You have stuff from Microsoft, IBM?

The strategy is to enable you to learn, build and deploy. We have the learn-and-build part available right now. Learn is what happens on the Learn platform, and build is what happens on the Lab platform. When you make an account with us, you get access to JupyterLab, and that’s where we provide you access to various Python environments focused on IBM’s Qiskit, Google Cirq and OpenFermion, Microsoft’s Q#, AWS’ Braket, and many more. These are all the environments ready to go. All we have to do is click Activate, and all the Python packages related to, let’s say, IBM’s Qiskit, you don’t have to install, and you can just start importing right away: Import Qiskit Aer or import Qiskit Ignis — all of those things are ready to go.

On top of that, one cool thing we do is, we clone all the tutorials available from IBM, Google and Microsoft into your home directly, which means you don’t have to Google search any of the tutorials. You just open that folder, and you have access to, let’s say “How to Do Finance” or “Portfolio Optimisation Using Qiskit.” You just open the notebook, the Python environment is already there, you can start running the code. In fact, this morning, I showed someone how to change the ticker of the SOC and run the code. That, essentially, is the idea, where we wanted to provide you the entire ecosystem at your fingertips — anywhere it’s possible to automate things, anywhere it’s possible to save you time, we want to do that.

What I thought was neat is, you can add packages, and if a new version of something comes out, you’ll still have access to the old one in case it breaks what you did — sort of a regression situation, so you can test and see how it works, because a lot of these packages become a little unwieldy. You change one thing, and then everything just blows up, or “No, you have to run this one specific version,” and then you didn’t realise that, and then you’re sunk. A lot of beginners will find that those headaches are lifted, so that was impressive, at least to me.

This is useful for people going it alone. You can have people at a company doing this. Then, I was thinking along the lines of community: There’s competition there — other platforms are trying to build community. The way you guys seem to handle that is through the blogs section — you can not only work on things, but you could also share them. Do you want to talk about that?

We’re paying a lot of attention to how to build a community around the tools that we’re building. When you log in to Lab, you’ll see that you have access to an extension called Code Snippets, and this was our attempt to build a mini Stack Overflow within JupyterLab. You can go in there, and you’ll see that there’s a code snippet in there. Copy it and paste within your notebook. As the community grows, we hope that a lot more people will contribute and other people will find it useful because you don’t have to go looking for that exact piece of code that will do the job on the internet or Google search if it could be embedded in your development environment as well.

As I mentioned, ultimately, we want to provide you the ability to deploy. The starting point of that is blogs. These blogs are an industry first — when you publish a Jupyter notebook as a blog, not only can your readers read and play with the interactive stuff, they can also change the code you have in there, and play with the code itself. The blogs are kernel-powered, which means when your readers come to our platform and they read your blog, we connect a Python kernel to it so all your code will run if they want it to.

When you talk about the kernels and everything, are there limitations to the free account, like what you can do, what you can have running, etc.?

As of now, no. We’re still working through whether we want to offer that as a premium tier. We definitely think that there’ll always be a free tier along with the premium tiers that we plan to come up with. The amount of stuff that we want to bundle in the free tier, we still haven’t quite settled on it yet, as we’re still going through the development process for our product, making it much better. And as that happens, we will have a good definition of what will be available as a free tier, hopefully, in the next four to six months.

Now is definitely the time to get in there and start playing.

Absolutely. The other thing is, we plan to come out with access to hardware as well, so if you’re part of our community, you may be one of the first ones to get paid access.

Right now, what do you access on the back end? Is it pure simulation still?

We will provide you enough compute to run 10 qubit circuits locally. By “locally,” I mean, every JupyterLab instance gets associated with a decent amount of compute resources. With those resources, you can run a simulation of 10 qubits, at least, and then you can connect to various hardware if you have separate accounts. What that means is, just like you would connect to IBM’s quantum computer or AWS’s quantum computer by putting in your credentials, you can do the same thing from qBraid. Eventually, we are working on a product where you wouldn’t even have to deal with that, where we would manage access to all those clouds. All you would need is to make a qBraid account and submit jobs from a central location, and specify whether you want to run on all the quantum computers available to AWS or IBM and so on. We would manage all the authentication and the costs associated with it.

Right now, do you have something in place with Braket from Amazon?

Yes. With Braket, we are working with the AWS team where, hopefully, in three to six months, you’ll see all those quantum computers becoming available on qBraid without you having an AWS account. That’s crucial. If you have an AWS account right now, or an IBM account or a Microsoft account, this is something that you can do right now. In fact, a lot of people who attended MIT’s quantum computing hackathon yesterday and on Saturday, they used qBraid to connect to IBM’s quantum computer and a quantum computer available through Quantum Inspire.

You’re going to be allowing customers to pass through even if they don’t have any account, which is super convenient, obviously. Do you anticipate any other way to draw junior developers in? Do you have any plans for publicising contests or anything like that of your own? I don’t know if anything else in the environment has inspired you in the community.

Absolutely. I think one of the things we’ve been thinking is, we want to do such an amazing job, that our platform should be so robust, that people wouldn’t feel the need to learn a lot of the skill sets associated with the Python environment. We want to get to a position where a quantum scientist or junior people starting in the field of quantum computing don’t ever feel the need to learn skill sets associated with quantum computing.

This is where the story is personal to me, where I was doing a PhD, and I realised that if I have to excel at writing amazing quantum code, that means I will have to be a master in Python. I asked myself, “Really? Why do I have to do that? If it could be possible that somebody handled this for me so I could focus on the quantum part, like running VQEs or coding of the exact algorithm, that would be nice.” This is where we want to provide that environment to all the future scientists or future developers, where they will come in and we manage all those things for them, and they focus on the quantum aspect of the problem.

You’re trying to move them way up the stack — just dealing with the algorithms, and then allow for the levels below to happen unseen. This will allow different back-end targets without thinking about it, will you be able to just code, essentially, and then transpile and send to different hardware targets ultimately?

Absolutely. I want to clarify: What we want to take away is the pains associated with Python and classical software. We want to provide users access to the entire quantum stack because we don’t think, right now, we’re there yet, where in quantum computing, you can access the high-level algorithm stack and not have to deal with the lower level.

We are working on something called qBraid-SDK. In fact, it is already public, and people can reach out to us and we provide them access to it. Through that, we have a transpiler built in where you can build your circuit in Qiskit, and we allow you to run it on AWS’s Braket, or Google’s quantum computer. If there’s a mismatch between the language you’ve written your code in and the quantum computer you specify you want to run your programme on, we transpile the circuit on the fly, run it on the hardware and get you the results.

This is where we do that, but this is not the most optimised transpilation, which I don’t think we can guarantee for the next five years, because of how much research that is going on in various aspects of it. If there’s somebody out there who picks out one teeny aspect, optimises it, and then we optimise a new road that was generic, it’s not valid. We want to keep making it better, and as the time goes by, you’ll see it keeps getting better. We will provide users the access to the quantum circuit itself, so if they feel the need, they can optimise it further.

I was wondering if, one day, we’re ever going to get to a point where “This machine type is better in this instance,” and then you do have to spend the time to drill all the way down and see if you can squeeze every little bit out.

That seems like a solid approach for now, especially with this whole idea of education and bringing people in. Do you have any plans for rewarding any users that come up with things that they share? Anything for the future to incentivise that community-building?

That’s a great question. We are still working on our incentives. We have something called qBraid Karma. It’s not active yet, but it will be a combination of the number of rewards you get for your code snippets and the number of upvotes you get for your blogs. We also have news, events and research papers on our platform. You can go in there and submit the upcoming events, and you can get votes on that. At some point, we had a proposal to take all the upvotes and do a modular square, and add them up to get the qBraid Karma. We haven’t settled on that. We may just go with the pure upvoting.

As if quantum computing wasn’t nerdy enough, you have to add that extra layer to it.

Yes. These are the things we want to keep doing to keep people entertained. While you’re developing, things should be fun, and we want to provide people the most fun platform — hide a lot of Easter eggs in there.

The platform’s already pretty slick-looking, but you had mentioned before we had this interview that you might even be doing a new UI. Were there any design elements of that you wanted to share? Anything that’s coming that’s exciting?

It’s mostly going to be a new design. Right now, the feel of the platform is a dark theme, and this is where we’ve heard feedback that people like light mode and it’s more welcoming, so that’s where we’re moving to. You’ll see the new UI is more welcoming. I don’t know if you saw our old website. It was also dark themed — the landing page. Right now, it’s light mode and seems a little more slick. We want to follow the same approach for Account.qBraid.com, and then, ultimately, at some point, we are working to redesign even the JupyterLab UI and make it more welcoming and a little more slick.

Right now, it still looks very much like a person comes in and creates an account. They could be from a company, of course, or they could just be a student. Do you have any plans for creating some kind of managed environment where, let’s say, you find a customer has 20 developers and they want to build some kind of coordination in there that’s private but still shareable? Do you plan on having any granular controls like that with accounts?

Absolutely. If you go and check out our environment manager, you’ll see we have a button for Create an Environment. This will let you create your own Python environment. You’ll be able to even start from, let’s say, IBM’s Qiskit as the base environment and install a bunch of stuff in there.

You can choose who goes in, like other users?

Exactly. Then, you’ll be able to share that exact environment with all your collaborators. This has always been one of the most crucial things for qBraid, where we wanted to provide you a platform where not only your code is reproducible, but also your runtime environment.

It’s interesting to scroll down the tree. Some of this might be hard to visualise unless you actually go to the site. I recommend you do, because it’s free and there’s really nothing stopping you. You can just go to qBraid.com and sign up, and of course there’ll be information in the show notes.

I think that pretty much covers everything. It’s always great to see approaches like this. Education is really important in this field right now. A lot of our customers ask us, “What do we do to train developers? What do we do to get them to the next level?” It’s nice to see everything all in one place, so I’m glad to see this approach.

Thank you for your kind words.

Thanks for joining us. I hope a whole bunch of new folks subscribe and join up after today.

Thank you very much.

Now, it’s time for Coherence, the quantum executive summary, where I take a moment to highlight some of the business impacts we’ve discussed today, in case things got too nerdy at times. Let’s recap.

qBraid is one of six next-gen startups in the nation’s first quantum accelerator, Duality. qBraid has come out with a cloud environment that lets users both learn and experiment with quantum programming. Based on Jupyter notebooks, the site imports tutorials from projects like IBM’s Qiskit and Microsoft’s Q# and gathers them in one place. Users can interact with the tutorials and write their own code, all without installing anything or worrying about dependencies.

While you’re learning on qBraid, the site lets you transpile code to run on any listed supported quantum hardware — for example, write in Google Cirq and run on Qiskit. If you have an account with one of the hardware providers or a cloud service like Amazon Braket, you can pass the code on to run on real machines. Otherwise, simulation of up to 10 qubits is also supported in the environment.

As users solve interesting problems, they’re encouraged to post blogs on the site. This doesn’t quite capture what gets posted. Blogs are actually interactive with code to help others learn. It would be interesting to see what kind of community grows here for knowledge sharing. Overall, the qBraid team is trying to create the easiest-to-use quantum coding platform, and it seems to be well on its way.

That does it for this episode. Thanks to Kanav Setia for joining to discuss qBraid. 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 Twitter and Instagram @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 ProtivitiTech on Twitter and LinkedIn. Until next time, be kind, and stay quantum curious.