Cloud ML Networking Engineer - EFA

Amazon; Seattle (WA); On-site; USD 110,500 - 160,000; Full time

Job summary

Amazon is seeking engineers to write high-performance C code for the Elastic Fabric Adapter (EFA) user-space stack, enabling hosts to network thousands of GPU/CPU instances for ML and HPC workloads.

You will contribute to Libfabric/Open MPI and collaborate across teams to design APIs, optimize occupancy, and support major AI customers. The role emphasizes hands-on performance work and scalable networking in the cloud.

Qualifications

• Bachelor's degree or above in computer science, computer engineering, or related field, or Bachelor's degree
• Strong programming skills in C (coursework, projects, or internship experience)
• Familiarity with systems programming concepts (memory management, performance optimization, Linux)
• Experience with source control (Git), code reviews, and build processes (internship experience counts)

Responsibilities

• You will write the highest-performing code in C for multiple open source projects supporting EFA, such as Libfabric and Open MPI.
• You will work with multiple teams in the stack to invent new APIs for the latest concepts in networking in the cloud.
• Dive deep into how your customers are doing collectives and messaging at high bandwidth and low latency.
• Provide expert-level support to some of the biggest names in AI in the world.

Skills

C programming

Linux

Git

Performance optimization

Systems programming

Education

Bachelor's degree in CS/CE

Tools

Git

Build systems

Libfabric/Open MPI familiarity

Job description

Amazon is seeking engineers to write high-performance C code for the Elastic Fabric Adapter (EFA) user-space stack, enabling hosts to network thousands of GPU/CPU instances for ML and HPC workloads.

You will contribute to Libfabric/Open MPI and collaborate across teams to design APIs, optimize occupancy, and support major AI customers. The role emphasizes hands-on performance work and scalable networking in the cloud.