z Performance Analyst

Introduction
At IBM work is more than a job – it’s a calling: To build. To design. To code. To consult. To think along with clients and sell. To make markets. To invent. To collaborate. Not just to do something better but to attempt things you’ve never thought possible. Are you ready to lead in this new era of technology and solve some of the world’s most challenging problems? If so lets talk.

Your Role and Responsibilities
The IBM Z Software Performance team is responsible for designing executing and analyzing stress workloads/benchmarks on the IBM Z and LinuxONE systems to ensure these systems meet customer expectations.
As a Software Performance Analyst/Tester you will design and create benchmarks and stress workloads execute performance measurements with those benchmarks and workloads analyze the performance measurements and guide Hardware Operating System and Software Development teams on performance improvements collaborating with a global IBM team.

• Design and implement benchmarks and stress workloads for AIU IO Card and keep them current
• Set up benchmarks and stress workloads and the underlying AIU IO Card configuration for different performances.
• Automate performance measurements and data collection for benchmarks and stress workloads
• Develop and enhance data collection and analysis tools
• Execute performance benchmarks and stress workload
• Analyze the performance measurements and collected data for performance issues and bottlenecks
• Guide Development teams across the stack (IBM Z Hardware/IBM Research IBM AIU application stack Middleware/Applications) on the fixing of performance issues due to configurations.
• USP : Working on Systems which consist of AIU IO cards which are developed for creating a platform for training and analysis of IBM Large Language Models for Generative AI.

Required Technical and Professional Expertise
As Individual Developer

• Overall experience of 6-10 years in performance measurement analysis and system testing skills
• Bachelors degree in Computer science Information Science.
• Basic ML/AI model architecture training and inferencing knowledge
• 3 years experience with Pytorch
• Source code repository systems (e.g. git) scripting language and test automation skills
• Linux administration basic skills
• Experience working with (docker/podman) containers
• Programming languages: Python C/C++ Bash
• English (fluent) language skills

Keywords: System Testing AND Performance analysis/measurement AND PyTorch AND DOcker/Podman AND Python c/c++ Bash AND Linux Skills

Preferred Technical and Professional Expertise
Master

• Know-how in AI transformer model design or modification
• Experience in TensorFlow and model inference serving (TF serving Nvidia Triton vLLM)
• Experience with hardware design and debugging skills
• Experience in performance profiling (Linux perf) and tracing
• Linux administration advanced skills
• AI accelerator hardware architecture knowledge (GPU TPU AMX)
• Programming languages: Cuda Java