AI Performance Engineering Intern: Optimize AI Systems

Hewlett Packard Enterprise Development LP; Town of Texas, Northern (WI, KY); Hybrid; USD 48,000 - 56,000; Full time

Benefits offered by this job

Health & Wellbeing

Personal & Professional Development

Unconditional Inclusion

Job summary

Hewlett Packard Enterprise Development LP is seeking an AI Performance Engineering Intern to design test plans and run experiments that quantify AI system performance. You will use profiling tools and work with GPUs, including CUDA and mixed-precision techniques, in Linux environments.

The role requires coursework in CS/Math/Software Eng and strong Python skills, plus familiarity with PyTorch/TensorFlow and AI hardware concepts.

Qualifications

• Working toward a B.S. in Computer Science, Applied Mathematics, Software Engineering or equivalent.
• Coursework and experience with GPU (NVIDIA/AMD) hardware architectures and parallel programming techniques, including parallel programming methods such as NCCL/RCCL, NVSHMEM, or OpenSHMEM
• Background with AI applications and their programming models/languages, including PyTorch/TensorFlow
• Proficiency in Python
• Deep understanding of LLM parallelization methods TP/PP/DP, and especially EP (MoE) is a plus
• Experience with using inference serving engines (e.g. vLLM, Ollama) is a plus
• Experience with AI clusters using job schedulers (e.g., PBS, Slurm) is a plus
• Experience on GPU-based AI clusters with Docker or Apptainer Containers is a plus
• Experience with GPU profiling using NVIDIA Nsight or AMD rocprof is a plus
• Proficiency with CUDA/ROCm is a plus
• Proficiency with the Linux operating systems, scripting, and knowledge of performance monitoring tools
• Experience with AI industry standard benchmarks (MLPerf) is a plus
• Self-driven, solid work ethic, and motivation
• Ability to multi-task with strong organizational skills and attention to detail
• Excellent analytical and problem-solving skills
• Integrity, customer focus, innovation, teamwork, and accountability in daily work
• Ability to work in a fast-paced, multitasking environment
• Excellent written and verbal communication skills. Mastery of the English language is required.

Responsibilities

• Work with team members to design and create test plans to understand an AI system's performance.
• Create and execute experiments to understand the performance characteristics of the system under test (SUT).
• Use profiling and/or monitoring tools to assist with the performance characterization of the SUT.
• Collaborate with team members regarding status, project progress, and issue resolution.

Skills

Python

CUDA

PyTorch/TensorFlow

GPU architectures

Linux

NCCL/RCCL

NVSHMEM

OpenSHMEM

Education

Working toward B.S. in Computer Science, Applied Mathematics, Software Engineering or equivalent

Tools

Docker

Apptainer

NVIDIA Nsight

AMD rocprof

PBS

Slurm

Job description

Hewlett Packard Enterprise Development LP is seeking an AI Performance Engineering Intern to design test plans and run experiments that quantify AI system performance. You will use profiling tools and work with GPUs, including CUDA and mixed-precision techniques, in Linux environments.

The role requires coursework in CS/Math/Software Eng and strong Python skills, plus familiarity with PyTorch/TensorFlow and AI hardware concepts.