AI/ML Distributed Training Engineer

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

Job summary

Amazon Inc. in Cupertino is seeking a Software Engineer- AI/ML to help build and optimize AWS Neuron Training for large-scale model training on Trainium. You will work across PyTorch and Neuron to implement distributed training components and apply parallelism strategies.

You will profile workloads to locate bottlenecks and collaborate with compiler, runtime, and libraries teams to land fixes, while building internal tooling and tests that ensure reproducibility.

Qualifications

• Bachelor's degree in CS or equivalent.
• 2+ years designing or architecting scalable systems.
• 3+ years full SDLC in production.
• 3+ years programming in Python or C/C++.
• Familiarity with LLM/transformer fundamentals.

Responsibilities

• Implement and tune components of distributed training stack for large-scale training.
• Contribute to data, tensor, and pipeline parallelism and apply reduced-precision formats.
• Profile workloads to identify bottlenecks across compute, memory, and host overhead.
• Build and maintain internal tooling, benchmarks, and tests for reproducible performance work.
• Take on increasing ownership as you grow in the role.

Skills

Python

C/C++

Distributed systems

LLM/transformers

PyTorch/JAX

Education

Bachelor's degree or above in CS or equivalent

Master's degree or above in CS or equivalent

Tools

PyTorch/JAX

Distributed libraries

RL frameworks

Job description

Amazon Inc. in Cupertino is seeking a Software Engineer- AI/ML to help build and optimize AWS Neuron Training for large-scale model training on Trainium. You will work across PyTorch and Neuron to implement distributed training components and apply parallelism strategies.

You will profile workloads to locate bottlenecks and collaborate with compiler, runtime, and libraries teams to land fixes, while building internal tooling and tests that ensure reproducibility.