- Hours
- Full time
- Where
- Upton, NY
- Posted
- Aug 17
The Scientific Computing Applications group in CSD has an immediate opening for a Postdoctoral Research Associate to design, develop, and deploy machine-learning and high-performance computing workflows, algorithms, and software in support of Department of Energy (DOE) mission applications across a broad range of scientific domains, including materials, biology, physics, and nuclear science. The successful candidate will partner closely with domain scientists to co-develop and apply cutting-edge machine learning and computational techniques to address scientific computing needs such as scalable ML training and inference, surrogate modeling of scientific processes, workflow automation and adaptive simulation pipelines, and performance analysis and optimization.
What you'd do
- Work collaboratively with computer scientists, computational scientists, applied mathematicians, and domain scientists
- Develop software that integrates machine learning and numerical techniques targeting heterogeneous architectures (GPUs and accelerators), including DOE leadership-class supercomputing facilities, to enable and support scientific research
- Identify and implement strategies for correctness and reproducibility testing, and analyze and optimize software scalability and performance
- Apply software engineering and documentation best practices to ensure usability and maintainability
- Present results at meetings, workshops, and conferences
- Publish findings in conference proceedings and/or peer-reviewed journals
- PhD in Computational Physics, Chemistry, Materials Science, Computer Science/Engineering, Applied Mathematics, or a related field.
- Strong experience developing, deploying, and optimizing applications and workflows in high-performance computing (HPC) environments.
- Demonstrated programming proficiency in C/C++ (preferred) and Python, with experience in additional languages such as Fortran considered a plus.
- Strong knowledge of at least one parallel programming model commonly used in HPC, such as MPI, OpenMP/OpenACC, CUDA, HIP, Kokkos, or SyCL/OpenCL.
About Brookhaven National Laboratory
Brookhaven National Laboratory (www.bnl.gov) delivers discovery science and transformative technology to power and secure the nation’s future. Brookhaven Lab is a multidisciplinary laboratory with seven Nobel Prize-winning discoveries, 37 R&D 100 Awards, and more than 70 years of pioneering research. The Lab is primarily supported by the U.S. Department of Energy’s (DOE) Office of Science. Brookhaven Science Associates (BSA) operates and manages the Laboratory for DOE.
- Starts
- 2026-08-06