PhD Intern - Machine Learning Performance Analyst

Overview

The Physical and Computational Sciences Directorate (PCSD) researchers lead major R&D efforts in experimental and theoretical interfacial chemistry, chemical analysis, high energy physics, interfacial catalysis, multifunctional materials, and integrated high-performance and data-intensive computing.

PCSD is PNNL’s primary steward for research supported by the Department of Energy’s Offices of Basic Energy Sciences, Advanced Scientific Computing Research, and Nuclear Physics, all within the Department of Energy's Office of Science.

Additionally, Directorate staff perform research and development for private industry and other government agencies, such as the Department of Defense and NASA. The Directorate's researchers are members of interdisciplinary teams tackling challenges of national importance that cut across all missions of the Department of Energy.

Responsibilities

The Future technology Computing group seeks PhD interns for the summer of 2025 with a strong background in distributed Machine learning training and inference, modern C/C++ programming, and cluster and job scheduling performance characterization and modeling. The duration of the internship is 3 months. The internship can be either remote or onsite based on the availability of the candidate. The candidate will be expected to use and familiarize themselves with world leading technologies which are available at the Pacific Northwest National Laboratory. Moreover, the candidate is expected to collaborate closely with domain scientists and computer scientists. The expected outcome involves high quality research work.
 
Responsibilities and Accountabilities

• Evaluating the performance of distributed machine learning models and frameworks
• Developing new performance models for AI/ML
• Designing, implementing, and evaluating new compiler optimizations for machine learning models and frameworks

Qualifications

Minimum Qualifications:

• Candidates must be currently enrolled/matriculated in a PhD program at an accredited college.
• Minimum GPA of 3.0 is required.

Preferred Qualifications:

• Strong parallel programming experience (e.g., Pthreads, OpenMP, MPI, Cuda)
• Background in compiler frameworks (e.g., MLIR, LLVM, TVM, XLA)
• Experience with distributed AI/ML frameworks (e.g., Tensorflow, Pytorch Distributed, Horovod)
• Experience in performance modeling
• Research experience preferred