Visiting Student - MCS - Garcia, Ainara - 8.21.26.

Argonne National Laboratory • Lemont, IL

12d • Onsite

About The Position

MPI4AI is a multi-university collaboration between Tennessee Technological University, the University of Tennessee, Knoxville, Stony Brook University, and the Illinois Institute of Technology. It aims to specialize MPI collective communication patterns for the needs of modern AI workloads running at scale on HPC systems, including training and inferences using various levels of parallelism (data, tensor, pipeline, sequence), agentic collaboration, neural architecture search. Aspects such as how to relax consistency (e.g. ignore stragglers), integrate with heterogeneous memory tiers (GPU HBM, host memory, SSDs, external storage) and how to enable asynchronous, GPU-centric communication are crucial in this context and will be advocated for the upcoming MPI-5 and MPI-6 standards. The student's role in the project will be to co-design communication patterns specifically tailored for AI applications (training, inference, agentic workflows) and to identify novel ways to apply them in AI runtimes (Torch, vLLM, DeepSpeed, etc).

Requirements

• The entirety of the appointment must be conducted within the United States.
• Must be 18 years or older at the time the appointment begins.
• Currently enrolled in undergraduate or graduate studies at an accredited institution.
• Graduated from an accredited institution within the past 3 months; or
• Actively enrolled in a graduate program at an accredited institution.

Responsibilities

• Co-design communication patterns specifically tailored for AI applications (training, inference, agentic workflows).
• Identify novel ways to apply communication patterns in AI runtimes (Torch, vLLM, DeepSpeed, etc).

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: No Education Listed