Senior ML Network Stack Engineer — Scale EC2 AI

Amazon Inc.; Seattle (WA); Sur place; USD 168 000 - 227 000; Plein temps

Résumé du poste

Annapurna Labs, an integral part of AWS, seeks a Lead Software Engineer to own the ML network stack for EC2 distributed AI/ML systems. You will lead across the ML communication stack, mentor engineers, and drive scalable features for the largest AI workloads.

You'll work with NCCL, NVSHMEM, NIXL, NCCL GIN and CUDA kernels, while applying Linux networking and HPC-style interconnects to optimize EC2 infrastructure for ML customers.

Qualifications

• 5+ years leading design or architecture of scalable systems.
• 5+ years full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
• Experience mentoring, tech lead or leading an engineering team, or experience in development in the last 3 years
• 5+ years of experience with programming language: C or C++

Responsabilités

• Lead across the NVIDIA ML communication stack for enabling NVIDIA GPUs to work with the AWS EC2 machines.
• Knowledge of ML Networking, ML applications, and frameworks will be highly regarded.
• You'll be leading senior, mid-level, and junior SDEs and directing work to ensure the team delivers functions and features required for the latest and largest ML workloads.
• Collaborate across teams and drive architecture to deliver scalable solutions for EC2 infrastructure.

Connaissances

Leadership

C/C++

Distributed systems

Linux networking

Mentorship

High-speed networking

Formation

Bachelor's degree in Computer Science or equivalent

Outils

NCCL

NVSHMEM

NIXL

NCCL GIN

CUDA

Description du poste

Annapurna Labs, an integral part of AWS, seeks a Lead Software Engineer to own the ML network stack for EC2 distributed AI/ML systems. You will lead across the ML communication stack, mentor engineers, and drive scalable features for the largest AI workloads.

You'll work with NCCL, NVSHMEM, NIXL, NCCL GIN and CUDA kernels, while applying Linux networking and HPC-style interconnects to optimize EC2 infrastructure for ML customers.