Software Development Manager, AWS Neuron SDK - Distributed Training
Core
Lead a team of engineers to design and deploy software for AWS Neuron SDK and Trainium-based servers, enabling distributed machine learning training at scale.
Role type
Senior IC Software Development Manager (ML Systems)
Builds
Production-ready distributed training integrations and tooling for PyTorch, JAX, and frontier model architectures (LLMs, MoE) on Trainium.
Domain
Cloud-scale Machine Learning / AI Infrastructure
Deliverable
production ML models
Required skills
Machine Learning, distributed training, PyTorch, JAX, system architecture, team management, performance optimization, LLMs, deep learning workflows
Preferred skills
LLM deployment on AI accelerators, CUDA kernels, low-level ML kernels, debugging large-scale systems, managing ML scientists
Technologies
PyTorch, JAX, Trainium, Neuron SDK, FSDP, DeepSpeed, Megatron
Responsibilities
Lead engineering team for Neuron SDK/Trainium platform; own customer onboarding from evaluation to production; drive initiatives to maximize Model FLOPS Utilization (MFU); build automation for time-to-first-training; partner with compiler/runtime teams; develop scalable processes for distributed training; drive technical strategy for frontier models; recruit and mentor ML systems engineers.
Seniority
Senior, hands-on IC with management responsibilities
