About the job
AWS Neuron is the complete software stack for the AWS Inferentia and Trainium cloud-scale machine learning accelerators and the Trainium-based servers that use them. As the SDM of Software Development for the Neuron Training team, you will be responsible for leading a strong team of engineers and managers to help design and deploy these new products. A successful candidate will have an established background in developing Machine Learning products with direct customer-facing experience, a strong technical ability and a motivation to achieve results. Experience in Machine Learning and software development is also a must.
Responsibilities
Lead a team of engineers focused on enabling new ML training customers on the Neuron SDK / Trainium platform.
Own the customer onboarding journey from model evaluation through production training at scale
Drive engineering initiatives to maximize Model FLOPS Utilization (MFU) for customer workloads through performance analysis, profiling, and tuning tools.
Build and maintain tooling, automation, and documentation that accelerates time-to-first-training for new customer models.
Partner with compiler, runtime, and framework teams to identify and resolve blockers in customer workloads.
Develop scalable processes for distributed training enablement (FSDP, DeepSpeed, Megatron, custom parallelism strategies).
Drive technical strategy for supporting frontier model architectures (LLMs, MoE, multi-modal) on Trainium.
Build strong cross-functional partnerships with product management, developer relations, and customer-facing teams.
Recruit, mentor, and grow a high-performing team of ML systems engineers.
Qualifications
Minimum
Experience working with PyTorch or JAX software
3+ years of engineering team management experience
7+ years of working directly within engineering teams experience
3+ years of designing or architecting (design patterns, reliability and scaling) of new and existing systems experience
Experience partnering with product or program management teams
3+ years of experience in deep learning / machine learning, including model training workflows
Experience with distributed training at scale (multi-node, multi-accelerator)
Preferred
Experience in developing and deploying LLMs in production on GPUs, Neuron, TPU or other AI acceleration hardware
Experience directly managing scientists or machine learning engineers
Experience debugging, profiling, and implementing best software engineering practices in large-scale systems
Experience in developing and deploying LLMs in production on GPUs, Neuron, TPU or other AI acceleration hardware, or experience with CUDA kernels or ML/low-level kernels
Experience with performance analysis, profiling, and optimization for deep learning training workloads