ML Systems Engineer - Distributed Training on Trainium

Amazon; Cupertino (CA); On-site; USD 165,000 - 224,000; Full time

Job summary

Amazon's Annapurna Labs team at AWS builds Neuron to accelerate DL on Trainium. The role focuses on optimizing distributed training throughput across the Neuron stack, working with PyTorch and JAX to push frontier-scale models.

You will own parallelism strategies, profile workloads, and translate findings into framework requirements, contributing upstream to OSS. The team emphasizes collaboration across hardware and software layers, with opportunities to shape AI acceleration tech across

Qualifications

• 3+ years of professional software development experience.
• Experience with distributed training systems and performance optimization.
• Familiarity with PyTorch or JAX and modern ML frameworks.

Responsibilities

• Lead efforts to optimize distributed training performance on Trainium.
• Own parallelism strategies across data, tensor, pipeline, and context.
• Profile end to end to identify bottlenecks and drive fixes with teams.

Skills

Software development

Distributed training

PyTorch

JAX

Performance tuning

Education

Bachelor's degree in CS

Tools

Python

CUDA

Job description

Amazon's Annapurna Labs team at AWS builds Neuron to accelerate DL on Trainium. The role focuses on optimizing distributed training throughput across the Neuron stack, working with PyTorch and JAX to push frontier-scale models.

You will own parallelism strategies, profile workloads, and translate findings into framework requirements, contributing upstream to OSS. The team emphasizes collaboration across hardware and software layers, with opportunities to shape AI acceleration tech across