Postdoctoral Research Associate
Core
Design, develop, and deploy machine-learning and high-performance computing workflows, algorithms, and software to support DOE mission applications in materials, biology, physics, and nuclear science.
Role type
Postdoctoral Research Associate (Scientific Computing)
Builds
Scalable ML training/inference pipelines, surrogate models, and automated simulation workflows for DOE leadership-class supercomputers.
Domain
Scientific computing, high-performance computing, machine learning, nuclear science, materials science.
Deliverable
production ML models | infrastructure
Required skills
PhD in Computational Physics/Chemistry/Materials Science/CS/Applied Math, HPC application development and optimization, C/C++ and Python programming, parallel programming (MPI, OpenMP, CUDA, HIP, Kokkos, SyCL), end-to-end ML model training and evaluation, software engineering best practices.
Preferred skills
Multi-node HPC ML scaling, ML architecture optimization for scientific/HPC, software performance profiling on CPUs/GPUs, numerical algorithms (linear solvers, MCMC), agentic computational workflows, exascale platform experience (Perlmutter, Frontier, Aurora), open-source contributions.
Technologies
MPI, OpenMP, OpenACC, CUDA, HIP, Kokkos, SyCL, OpenCL, Python, C/C++, Fortran, GPUs, supercomputers (Perlmutter, Frontier, Aurora).
Responsibilities
Develop software integrating ML and numerical techniques for heterogeneous architectures; implement correctness/reproducibility testing and optimize scalability; present results at conferences; publish in peer-reviewed journals.
Seniority
Postdoctoral Researcher (early-career IC)