AI/ML Software Engineer - Trainium Distributed Training

Amazon; Seattle (WA); On-site; USD 144,000 - 194,000; Full time

Job summary

Amazon Neuron Training team in Seattle, WA, part of AWS, builds the Neuron SDK to accelerate DL/GenAI workloads on Trainium. We design hardware-software co-design solutions spanning frameworks, kernels, compiler, runtime and collectives to maximize training throughput.

The role focuses on leading optimization efforts for distributed training, working across PyTorch and Neuron stacks, defining parallelism strategies, and profiling workloads to push performance limits.

Qualifications

• 3+ years of non-internship professional software development experience.
• 2+ years of non-internship design or architecture of systems.
• Experience programming with at least one software language.

Responsibilities

• Lead efforts to optimize distributed training performance on Trainium.
• Collaborate across PyTorch and Neuron software stack to enable large-scale training.
• Own parallelism strategies across data, tensor, pipeline, expert, and context.
• Profile workloads to identify compute, memory, collectives, or host bottlenecks and drive fixes.
• Translate performance gaps into framework requirements and contribute upstream.

Skills

Non-internship software development

Software design/architecture

Programming experience

Education

Bachelor's degree in Computer Science or equivalent

Job description

Amazon Neuron Training team in Seattle, WA, part of AWS, builds the Neuron SDK to accelerate DL/GenAI workloads on Trainium. We design hardware-software co-design solutions spanning frameworks, kernels, compiler, runtime and collectives to maximize training throughput.

The role focuses on leading optimization efforts for distributed training, working across PyTorch and Neuron stacks, defining parallelism strategies, and profiling workloads to push performance limits.