Principal AI/ML HPC Architect & TAM - Enterprise Support

Amazon Web Services (AWS); Dallas (TX); On-site; USD 183,000 - 247,000; Full time

Job summary

Amazon Web Services, Inc. seeks a Principal AI/ML HPC Specialist to guide NAMER Research labs and enterprise customers through advanced AI/ML transformations. You’ll architect GPU/Trainium clusters, optimize NCCL communications, and ensure scalable, cost‑effective training and inference at scale.

You will collaborate with TAMs, SAs, and service teams to deliver best‑practice guidance, HPC to AI convergence patterns, and enterprise‑grade solutions that drive operational excellence.

Qualifications

• Bachelor's degree in a related field.
• 8+ years of AI/ML, distributed computing, or GPU‑accelerated infra.
• 3+ years designing, implementing, or consulting on large‑scale ML training/inference for customers.
• 10+ years IT development or implementation/consulting in software, cloud, or AI/ML.
• Experience with PyTorch, TensorFlow, or JAX in production or research.

Responsibilities

• Lead strategic technical engagements for enterprise AI/ML workloads.
• Design production‑grade AI/ML training and inference solutions with Slurm and distributed frameworks.
• Enable customer success with HPC capabilities and MLOps pipelines.
• Partner with service teams to improve training throughput and GPU utilization.

Skills

AI/ML

Distributed computing

GPU infrastructure

Education

Bachelor's degree

Tools

PyTorch

TensorFlow

JAX

Slurm

NCCL

SageMaker HyperPod

AWS PCS/ParallelCluster

Job description

Amazon Web Services, Inc. seeks a Principal AI/ML HPC Specialist to guide NAMER Research labs and enterprise customers through advanced AI/ML transformations. You’ll architect GPU/Trainium clusters, optimize NCCL communications, and ensure scalable, cost‑effective training and inference at scale.

You will collaborate with TAMs, SAs, and service teams to deliver best‑practice guidance, HPC to AI convergence patterns, and enterprise‑grade solutions that drive operational excellence.