Postdoctoral Research Associate - HPC System Architecture and Storage Design

Postdoctoral Research Associate - HPC System Architecture and Intelligent Storage Design

The National Center for Computational Sciences (NCCS) at the Oak Ridge National Laboratory (ORNL) is seeking a postdoctoral research associate in High-Performance Computing (HPC) system architecture and intelligent storage design. The candidate will contribute to research and development efforts in scalable storage and memory architectures, telemetry-driven system optimization, and application-driven performance analysis for HPC, scientific Artificial Intelligence (AI), and scientific edge computing.

Specific areas of research interest include: telemetry collection and analysis to inform system optimization, workload characterization, and predictive fault tolerance in HPC systems. Architectural exploration and performance modeling of high-bandwidth memory (HBM) and DDR memory systems in the context of data-intensive scientific computing. Redesign of storage systems to meet evolving demands in AI/ML and edge-to-HPC workflows, including support for data movement, retention policies, and user-defined storage behaviors.

Major duties and responsibilities include: collaborating with internal and external researchers on a variety of data and storage-related research projects for use cases in HPC, scientific AI, and scientific edge computing; conducting I/O and storage performance characterization of HPC and scientific AI applications or libraries on multi-tier HPC storage systems; collecting, analyzing, and leveraging telemetry data from HPC systems to support data-driven design decisions and performance tuning;

exploring trade-offs in DDR and HBM usage patterns, and proposing architectural improvements for mixed-memory and bandwidth-intensive workloads; designing and evaluation of ephemeral, user-configurable, and composable data and storage systems; designing system-level approaches for time-sensitive or data-intensive processing of data originating at scientific edge systems using large-scale HPC/AI computational and storage systems; coauthoring peer-reviewed publications, technical reports, and presentations; seeking membership and service opportunities in professional, academic, and research organizations.

Basic qualifications include: a PhD in computer science/engineering or relevant area with an education and a research track record in HPC/AI/edge systems and storage research within the past five years. Preferred qualifications include: ability to work independently to design and deploy methods at scale; familiarity with hardware-software co-design, memory hierarchies (DDR, HBM), and system-level telemetry tools; experience in HPC and associated software development for applications, middleware, and/or system software;

flexibility to adapt to diverse R&D projects and tasks; effective communicator in both verbal and written forms; ability to collaborate with scientists, engineers, and sponsors; interest in mentoring student internships.

The appointment length will be up to 24 months with the potential for extension. Initial appointments and extensions are subject to performance and availability of funding.