AI Performance Analyst

Introduction

At IBM Infrastructure & Technology, we design and operate the systems that keep the world running. From high-resiliency mainframes and hybrid cloud platforms to networking, automation, and site reliability. Our teams ensure the performance, security, and scalability that clients and industries depend on every day. Working in Infrastructure & Technology means tackling complex challenges with curiosity and collaboration. You'll work with diverse technologies and colleagues worldwide to deliver resilient, future-ready solutions that power innovation. With continuous learning, career growth, and a supportive culture, IBM provides the opportunities to build expertise and shape the infrastructure that drives progress.

TheIBM Z Software Performance teamis responsible for designing, executing, and analyzing stress workloads and benchmarks onIBM ZandLinuxONE systemsto ensure these platforms meet stringent customer expectations for reliability, scalability, and performance.

The team develops and maintains a suite of tools and automation scripts to support:

• Performance environment setup and configuration
• Performance measurements and data capture
• Data storage in centralized repositories
• Presentation and visualization of performance metrics
• Analysis and comparison of captured data across multiple scenarios

Your role and responsibilities

Your Role and Responsibilities*

As a AI Performance Analyst you will be responsible for:

• Design and implement benchmarksand stress workloads forAIU IO Cards, ensuring they remain current and relevant.
• Set up benchmarks and stress workloads, including configuring the underlyingAIU IO Cardfor various performance scenarios.
• Automate performance measurementsand streamline data collection processes for benchmarks and stress workloads.
• Develop and enhance data collection and analysis toolsto improve efficiency and accuracy.
• Execute performance benchmarksand stress workloads to validate system performance.
• Analyze performance measurementsand collected data to identify bottlenecks and resolve performance issues.
• Collaborate with development teamsacross the stack (IBM Z Hardware, IBM Research, IBM AIU application stack, Middleware/Applications) to guide and support performance optimization efforts related to configurations.

Required education
Bachelor's Degree

Preferred education
Master's Degree

Required technical and professional expertise

• Overall Experience:5-10 years inperformance measurement,analysis, andsystem testing.
• Education:Bachelor's degree in Computer Science or Information Science.

Technical Skills

AI/ML Knowledge:

Basic understanding ofML/AI model architecture, training, and inferencing.

3+ years of experience withPyTorch, Tensorflow, vLLM

Development & Automation:

Proficiency insource code repository systems(e.g., Git).

Strongscriptingandtest automationskills.

System & Containerization:

BasicLinux administrationskills.

Hands-on experience withDockerandPodmancontainers.

Solid understanding ofOperating System fundamentalsandComputer Architecture concepts.

Basic experience inperformance analysisof applications/systems and familiarity withperformance tools.

Hands-on experience infunctionalandperformance testingof multi-tiered applications.

Additional Skills:

Exposure toAgile methodologiesand ability to apply agile concepts effectively.

Strongpresentation skillsand ability to communicate technical concepts clearly.

Collaborative team player with excellent interpersonal and communication skills.

Programming Languages and scripts:

Python,C/C++,Bash,Ansible, Java

Preferred technical and professional experience

• Master's degreein Information Technology, Computer Science, or Computer Engineering.
• AI/ML & Model Serving
  • Know‑how inTransformer model design and modification(architecture tuning, fine‑tuning, optimization).
  • Hands-on experience withTensorFlowand model inference serving usingTensorFlow Serving,NVIDIA Triton Inference Server, andvLLM.
• Systems Performance & Observability
  • Proficient inperformance profilingand tracing (e.g.,Linux perf, flame graphs, instrumentation).
  • AdvancedLinux administrationskills (networking, storage, process management, kernel parameters, automation).
• Hardware & Accelerators
  • Experience inhardware designanddebugging(board bring-up, driver interactions, performance counters).
  • Working knowledge ofAI accelerator architectures-GPU,TPU,AMX(capabilities, memory hierarchies, scheduling/tiling considerations).
• Programming Languages
• Proficiency inCUDA(kernel development, memory optimization, streams/concurrency) andJava(services, tooling, SDK integration).

Years of Experience:
3-10