Superconducting qubits take many forms. The one at the heart of IBM quantum processors, the transmon qubit, is there partly because of Hanhee Paik.

By figuring out how to improve the transmon qubit’s coherence time, Paik showed the world that a superconducting quantum computer was possible. IBM now operates a fleet of quantum computers worldwide based on her transmon qubit design. Today it’s Paik’s job at IBM to help researchers find innovative ways to use them.

She currently oversees several projects in Chicago, including the National Quantum Algorithm Center (NQAC) that IBM recently established with the state of Illinois, and whose members include the University of Illinois Urbana-Champaign (UIUC) and the University of Chicago.

IBM just signed a new agreement with UIUC to expand the Discovery Accelerator Institute, with the goal of advancing quantum-centric computing and developing algorithms that combine the strengths of HPC and quantum computing to achieve transformational results.

IBM will give UIUC students and researchers access to IBM Quantum computers alongside UIUC’s National Center for Supercomputing Applications (NCSA) Delta and DeltaAI supercomputers. This “quantum-centric” architecture will allow researchers to test new algorithms and explore quantum use cases in the real world.

Paik will help drive the agenda, drawing on her deep expertise in quantum systems and her growing knowledge of high-performance computing, picked up on a recent assignment in Japan.

We recently caught up with her to talk about IBM’s collaboration with UIUC, and the past, present, and future of quantum research.

How did you get involved in the UIUC collaboration?

I currently co-lead IBM Research’s academic collaboration program, and UIUC is one of our biggest collaborators. I'm also responsible for quantum initiatives across Illinois. That’s why I just moved to Chicago. In addition to this Institute renewal, IBM recently partnered with the Illinois Quantum and Microelectronics Park (IQMP) to create the NQAC and will bring an IBM Quantum System Two to Chicago in September. Governor Pritzker’s vision is to apply quantum to scientific and industry problems and grow the economy, and algorithms will be essential. Without algorithms, we can’t really use quantum computing to solve problems and develop the economy. We hope the center will be a place where researchers from across the state gather.

What will the next five years at the Discovery Accelerator Institute look like?

We want to advance the next chapter of computing. Our vision for quantum-centric supercomputing is to integrate CPUs, GPUs, and QPUs. We also want to advance AI platforms to handle the next generation of AI workloads. UIUC has strong expertise in scientific computing, and starting this year, they can access our IBM Quantum computers over the cloud to work alongside their HPC. One of our research goals is to efficiently integrate quantum and HPC workflows and make this technology accessible to students and researchers.

You make quantum-centric supercomputing sound easy!

It’s not. Sometimes you bring QPUs, sometimes GPUs and CPUs. The workflows are complex in an integrated system. Together with RIKEN, IBM pioneered the workflow to manage the quantum computational resources and execute the algorithms more efficiently. When you're doing a quantum computation, you can’t leave the HPC idle because thousands of people are using those systems together. They’re shared assets.

Is there one problem you’d like to see quantum centric supercomputing tackle?

I’m curious to see if an integrated quantum and GPU system can help us discover quantum-plus-AI algorithms. We’d like to discover more algorithms that can use both resources in an efficient way. The sample-based quantum diagonalization (SQD) algorithm is an example of an algorithm that uses the best of both quantum and classical. The [iron-sulfur] molecule, Fe4S4 is a key molecule that produces energy in your cells’ mitochondria through electron transfer. You can’t compute its ground state with brute force HPC. However, IBM and RIKEN researchers solved it using a quantum-centric supercomputer. This is a powerful example of the effectiveness of quantum-centric supercomputing algorithms.

What kinds of algorithms and applications will IBM and UIUC focus on?

Materials and condensed matter physics are key topics, and we plan to research algorithms that can solve properties of open quantum systems, systems with defects, and strongly correlated systems, which can behave like superconductors. The laws of physics explain how materials work, but we don’t fully understand the mechanisms that give rise to their properties. We hope that quantum-centric supercomputing can help us better understand the physics so we can discover new materials.

What new AI projects will you focus on?

We’d like to use AI to design more efficient chips and systems, a project we’re calling “algorithms to silicon to system.” This is not just about using AI for CPU designs. We’d also like to use AI to design QPUs, which need highly complex connectivity among qubits to run more efficient error correction code.

What stands in the way of wider quantum adoption?

I think one barrier is the perception that you need to learn quantum mechanics to work on quantum computing. We will continue to build abstraction layers that make quantum computers even more accessible. For example, IBM’s open-source quantum programming language, Qiskit, is designed to make quantum computing easy for everyone.

What have you learned from your friends in computer science?

I’ve learned a lot about HPC, software, and computing architectures and I’ve come to appreciate how important job scheduling is for managing shared computing resources. We need to think about data workflows to make the compute more efficient and how to manage quantum and HPC resources properly for seamless access. Computing languages are essential for making these systems more accessible to users without too much extra learning effort.

You went from designing qubits to thinking about data workflows!

We’re talking about quantum computing right now, but quantum data is only on the processor. Once it leaves the processor, it’s all zeros and ones that still contain “quantum information” processed by the quantum computer. We need to use classical data processed by our quantum processor to make it useful, and to do that, we need to think about the flow of zeros and ones. This is what I learned from hanging out with HPC scientists in Japan for two years!

IBM put the first quantum computer in the cloud a decade ago — it had just five qubits. The IBM Quantum System Two coming to Chicago will have more than 100 qubits. Did you see this day coming?

I’ve always felt like Alice, in Alice in Wonderland, chasing the bunny to the next place. I focused on one goal, and then the next, and the next, until I ended up where I am right now. If the field of quantum computing wasn’t progressing, then I probably would have stopped, because the bunny would have stopped. There’s always been something interesting to pursue.

I wanted to explore quantum algorithms and use one of the quantum computers I’d spent my entire career trying to build. But it wasn’t an easy pivot. As a hardware scientist, I didn’t know much about algorithms back in 2018. Jay Gambetta became a VP of IBM Quantum in 2019, and he was a quantum algorithms expert. When he offered me the position as his chief of staff I said yes thinking I could learn more about algorithms.

What did you do next?

After my tenure, Jay encouraged me to go into technical business development where I learned business skills and signed a few quantum business deals. I worked with legal teams and contract teams and watched my colleagues negotiate. It’s an art! After that, I went to Japan to help build the quantum team. I worked with them on quantum-centric computing software and algorithms and got to learn a lot of computer science. Now I’m in Chicago getting used to my new role leading IBM’s quantum algorithms partnerships.

How did you learn to navigate so many worlds — quantum, HPC, sales, academia?

I'm originally from South Korea and studied at Yonsei University which is now a client. When I moved to Maryland for my PhD, I could barely speak English or understand American culture, so I had to go through a long process of socialization. As an American now, I feel like a millennial because those were my peers in graduate school where I was resocialized. This resocialization experience taught me to be more flexible and adaptable. Today, I understand American culture, Korean culture, and now Japanese culture. When I went to work for Jay [now Research Director], I had to organize technical meetings, help manage teams, build narratives and presentations. I learned how to talk to a lot of different people. It was a tremendous growth experience.

Did you ever dream you'd be where you are now?

I never thought that far ahead. I was chasing the bunny. For my thesis, I worked on improving coherence times. It was an important topic for superconducting qubits because many people thought there was no future there. But if you could make coherence time longer, we could do computation. I moved to Yale for a second postdoc because I heard they had qubits with coherence times of 1.5 microseconds — much better than my thesis. So, I wanted to study at Yale, of course! At Yale, I designed a qubit with a 100 times improvement. That's when IBM reached out. They were already thinking about how to make quantum a business.