Birds of a Feather Archive

Continuum Computing: Agentic AI for Intelligent Orchestration


Authors: Neena Imam (Southern Methodist University), Nagi Rao (Oak Ridge National Laboratory (ORNL)), Kyle Chard (University of Chicago), Ravinder Kapoor (U.S. Department of Energy)

Abstract: Continuum computing (digital continuum) is a distributed ecosystem spanning edge devices, scientific instruments, cloud platforms, data centers, supercomputers, and quantum systems. As these infrastructures become increasingly complex, agentic AI offers new opportunities for intelligent orchestration of computation, data, workflows, and heterogeneous resources. This BoF will explore how collaborative AI agents can enable autonomous scientific computing while ensuring trust, resilience, and human oversight. Topics include multi-agent coordination, AI-driven resource management, autonomous workflows, foundation models for scientific computing, and new programming models. The session will identify key research challenges and foster cross-disciplinary collaboration to advance the next generation of intelligent continuum computing.

Long Description: The new paradigm of Continuum Computing (also known as the digital continuum) is a distributed, multi-layered ecosystem that spans sensors at the edge, interconnected instruments, cloud platforms, data centers, exascale supercomputers, and, more recently, quantum computers. Continuum computing has evolved to keep pace with the growth and expansion of geographically distributed scientific and computing infrastructures. In the continuum paradigm, computation and data are orchestrated across multiple stages to optimize data movement, resource utilization, and latency. Realizing this vision requires novel approaches to system design, software frameworks, adaptive workflows capable of responding to dynamic data, monitoring tools, multisite governance models, and meaningful experimental metrics. While advances in AI, quantum computing, and resilient architectures have significantly expanded the capabilities of the continuum, they have also introduced unprecedented challenges in managing increasingly dynamic, large-scale, and highly distributed computational environments.

We organized the inaugural Continuum Computing BOF at SC22, introducing the digital continuum and discussing the state-of-the-art. The SC23 BOF explored the technologies needed to enable the continuum through the convergence of previously distinct paradigms, including HPC, AI/ML, and digital twins. At SC24, we examined the role of quantum information science in extending the continuum beyond classical computing. The SC25 BOF focused on resilience, addressing the challenges of building a secure continuum.

For SC26, we propose the fifth BOF in this series to examine one of the most transformative developments in artificial intelligence—Agentic AI. As continuum computing grows in scale and complexity, the fundamental question shifts from how to build and secure the continuum to how to intelligently orchestrate it. Can collaborative AI agents become the operating model for future continuum computing, autonomously coordinating computation, data, workflows, and heterogeneous resources while maintaining human oversight? This BOF will explore the opportunities, challenges, and research directions for Agentic AI as the foundation for intelligent orchestration across the digital continuum. SC26 BOF topics will include multi-agent coordination, AI-driven resource management and scheduling, autonomous scientific workflows, new programming models, foundation models for scientific computing, human-agent collaboration, trust, governance, resilience, and oversight. We will discuss the research challenges, systems architectures, and community priorities needed to realize the next generation of intelligent continuum computing while addressing fundamental questions surrounding scalability, interoperability, transparency, and the evolving role of humans in increasingly autonomous continuum environments.

Attendees will receive an overview of the emerging opportunities and challenges associated with Agentic AI for intelligent orchestration across the continuum, along with a community-driven set of open research questions. This BOF aims to catalyze cross-disciplinary collaboration among systems researchers, AI scientists, networking experts, cybersecurity specialists, and application scientists to define the architectures, programming models, governance frameworks, and operational principles needed for the next generation of intelligent continuum computing. Given the rapid advances in Agentic AI and the growing complexity of distributed scientific infrastructures, this timely and interdisciplinary discussion is expected to generate significant interest within the SC community. The session leaders will provide complementary perspectives from academia (Imam), national laboratory (Rao, Chard), industry (TBD), and government (TBD).

Website: https://www.smu.edu/provost/odonnell-institute/news/sc26-bof



Back to Birds of a Feather Archive Listing