Senior GPU & AI Systems Engineer - Multi-Node Training

Micron Memory Malaysia Sdn Bhd; Boise (ID); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Medical benefits

Paid time off

401(k) plan

Job summary

Micron Technology's Boise, Idaho team seeks a seasoned GPU Performance Engineer to architect and optimize large-scale AI workloads on multi-GPU clusters. You'll train, fine-tune, and deploy models (SFT, RLHF) and push performance through distributed training and memory-efficiency techniques.

You will collaborate with data scientists, hardware architects, and software engineers, mentoring peers while advancing CUDA/HIP kernels, ML frameworks, and GenAI agent strategies.

Qualifications

• 10+ years in GPU hardware and software optimization across cloud and on‑prem.
• 5+ years in performance optimization and low‑level systems using C++ and CUDA.
• Deep proficiency with LLMs, prompt tooling, and PEFT methods (LoRA/QLoRA).
• Experience building scalable ML systems with distributed training (DDP, FSDP).
• Proficient in Python (preferred) or Java and CI/CD with Docker/Kubernetes.

Responsibilities

• Architect and execute large-scale model training and fine-tuning on multi-node, multi-GPU clusters.
• Optimize throughput and memory using distributed strategies and mixed precision.
• Design autonomous AI agents for multi-step reasoning and tool execution.
• Profile workloads to identify bottlenecks in compute, memory, and latency.
• Write high-performance kernels (CUDA/HIP) and collaborate with hardware architects.
• Mentor engineers in parallel programming and optimization techniques.

Skills

GPU architecture

Performance optimization

Distributed training

Python

C++

LLMs and PEFT

GenAI / AI agents

CI/CD tooling

Docker & Kubernetes

Communication skills

Education

Bachelor’s or Master’s in CS/Statistics or related

PhD preferred

Tools

CUDA

Triton

PyTorch

Kubernetes

Ray

Kubeflow

TensorRT-LLM

Job description

Micron Technology's Boise, Idaho team seeks a seasoned GPU Performance Engineer to architect and optimize large-scale AI workloads on multi-GPU clusters. You'll train, fine-tune, and deploy models (SFT, RLHF) and push performance through distributed training and memory-efficiency techniques.

You will collaborate with data scientists, hardware architects, and software engineers, mentoring peers while advancing CUDA/HIP kernels, ML frameworks, and GenAI agent strategies.