Amazon SDM ML Acceleration Neuron Inference Apps 

United States, California, Cupertino 

418178454

AWS Neuron is the complete software stack for the AWS Inferentia and Trainium cloud-scale machine
learning accelerators and the Trn1 and Inf1 servers that use them.As the Software Development Manager in the Neuron Inference team, you will be responsible for leading a strong team of engineers to help design and deploy ML applications/usecases on various frameworks such as Pytorch, JAX, Tensorflow. You will be responsible for the full development life cycle of our integrations and extensions for inference in Pytorch/JAX using XLA. Responsible for development, enablement and performance optimization (latency, throughput) of a wide variety of ML model families, including massive scale LLMs like Llama3, GPT2, GPT3 and beyond, as well as stable diffusion, Vision Transformers and many more. Deepspeed and other distributed inference libraries are central to this and extending all of this for the Neuron based system is key. Develop reliability/scalability features and performance updates in the Neuron ML Frameworks as well as contribute to other popular open Frameworks to enable them make Trainium and Inferentia devices as the first-class citizens for ML Acceleration. Lead the way to ensure support for key ML functionality in a combined chip / software platform. Ensure the right thing is being built and delivered to customersA successful candidate will have an established background in developing ML Inference features and performance optimizations using Pytorch/JAX on XLA devices. The ideal candidate should have a strong technical ability to work/deliver on a vertically integrated system stack that consists of a combinatorial matrix of hardware, frameworks, and workflows. Deep expertise in Framework integrations and development using C++ is a must along-with direct customer-facing experience and a strong motivation to achieve results.A day in the life
You will work with the executive leadership and other senior management and technical leaders to define product directions and deliver them to customers. We build massive-scale distributed training and inference solutions. This organization builds the full stack of software, servers and chips to accelerate at the highest scale.
Work/Life Balance
Mentorship & Career Growth

- 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

- 2 years experience leading teams that in Machine Learning development including building and training large models, working with Pytroch and/or Tensorflow using large distributed fleets of GPU or other accelerated systems.