Hybrid quantum-classical machine learning represents a frontier in computational research, combining the potential advantages of quantum computing with established classical optimization techniques. PennyLane provides a Python framework that seamlessly bridges quantum circuits and classical machine learning, enabling researchers to build, optimize, and deploy variational quantum algorithms. This paper introduces PennyLane as a versatile tool for quantum machine learning, optimization, and quantum chemistry applications. We demonstrate use cases including quantum kernel methods, variational quantum eigensolvers, portfolio optimization, and integration with classical ML frameworks such as PyTorch, TensorFlow, and JAX. Through concrete Python examples with widely used libraries such as scikit-learn, pandas, and matplotlib, we show how PennyLane facilitates efficient quantum circuit construction, automatic differentiation, and hybrid optimization workflows. By situating PennyLane within the broader context of quantum computing and machine learning, we highlight its role as a methodological building block for quantum-enhanced data science. Our goal is to provide researchers and practitioners with a concise reference that bridges foundational quantum computing concepts and applied machine learning practice, making PennyLane a default citation for hybrid quantum-classical workflows in Python-based research.
Quantum computing is entering the mainstream as part of cloud offerings, where it serves as a special-purpose accelerator in larger applications. However, it is still challenging for developers and researchers to design, build and manage the resources for Hybrid Quantum-Classical (HQC) applications that include both classical logic (x86, ARM) and quantum circuit blocks, and run across both traditional and quantum processors. Further, quantum hardware is available on public cloud and even on-premise as private clouds, with varying capabilities, costs and queue times. Further, such quantum circuits also expose optimization methods that offer cost, time, accuracy and parallelism trade-offs. So, there is a compelling for easy composition of HQC applications that can be effortlessly and efficiently deployed on hybrid clouds. In this paper, we propose a framework to intuitively compose and deploy “zero-touch” quantum-classical Function-as-a-Service (FaaS) workflows through various workflow patterns that leverage diverse cloud system (workflow partitioning, adaptive polling) and quantum (circuit cutting, qubit reuse) optimizations. These utilize our XFaaS FaaS workflow framework for hybrid cloud deployments on AWS and Azure, and IBM Qiskit SDK for the quantum circuit toolchain. We also offer detailed experimental profiling of these optimizations for realistic and synthetic HQC applications on real clouds, and on real and simulated quantum hardware, and analyze the benefits of cloud system and quantum circuit optimizations. Our results demonstrate up to 53 % improvement in time and 80 % in cost when quantum circuit optimization on hardware is used in conjunction with dynamic fan-out.
As quantum computing (QC) technologies mature, their integration into established high-performance computing (HPC) infrastructures is becoming a central objective for next-generation computing systems. However, unlocking the potential of hybrid platforms for computationally demanding workloads remains challenging. The mismatch between quantum and classical programming models, the limited maturity of quantum software stacks, and the scarcity of quantum processing units (QPUs) above all, necessitate scheduling strategies that go beyond standard HPC mechanisms to manage such heterogeneous and constrained resources.
To address this issue, we investigate three distinct methodologies for HPC-QC resource scheduling: time-based multiplexing, dynamic resource management, and workflow decomposition. Experimental validation on production HPC clusters and real quantum hardware demonstrates the effectiveness of these approaches under different workload scenarios. Malleability and workflow strategies significantly optimize classical resource utilization, reducing consumption by up to 45.7% and 64% respectively, proving to be best fitted for hybrid jobs where quantum and classical workloads are evenly balanced. Conversely, time-multiplexing enhances QPU utilization and reduces execution time at the cluster level, making it the optimal strategy for the opposite context, which is characterized by high classical-quantum workload imbalances. These findings underscore the practical viability of tailored scheduling strategies for hybrid HPC-QC environments and highlight their complementarity in building efficient, scalable software stacks for next-generation quantum-accelerated facilities.
Scientific workflows constitute a pivotal resource for the scientific computing community, as they provide a standard interface to define, execute, and analyze the results of scientific computations. The advent of the Post-Moore era exposes the scientific computing research community with the challenge of scaling HPC facilities, to address the demands of modern scientific applications. While quantum computing, on the one hand, promises noteworthy theoretical speedups for different scientific applications, executing a whole scientific workflow on quantum hardware is infeasible since not all computations can achieve a speedup on quantum hardware. As a consequence, computational scientists defined the concept of a hybrid quantum-classical workflow, which is capable of exploiting the capabilities of quantum and classical hardware. Since the adaptation of classical scientific workflows to the new hybrid quantum-classical scenario can be challenging for scientists, we propose RIGOLETTO, a workflow definition language allowing scientists to define where to execute each task on a hybrid quantum-classical system. As a proof-of-concept, we report on two scientific workflows and identify open challenges in the definition of hybrid quantum-classical workflows.