Software Engineer, Trainium Collectives

Requirements:

• We need 3+ years of non-internship professional software development experience.
• We need 2+ years of non-internship experience designing or architecting new and existing systems, including design patterns, reliability, and scaling.
• We need programming experience in at least one software language.
• We prefer 3+ years of full software development life cycle experience, including coding standards, code reviews, source control management, build processes, testing, and operations.
• We prefer a bachelors degree in computer science or an equivalent background.

Responsibilities:

• We enhance collective algorithms and topologies to improve training performance.
• We use tools such as Neuron Explorer to uncover bottlenecks in compute and bus bandwidth utilization.
• We monitor and evaluate processor, DMA, firmware, and workload metrics.
• We optimize collective operations to scale AI compute across the data center through low-level device driver development.
• We collaborate closely with the hardware team to co-optimize software and Trainium silicon.
• We develop and improve C/C++ implementations of collective communication patterns.
• We investigate and apply enhancements for training topologies used by modern LLMs.
• We build and maintain analysis frameworks and automation solutions.

Technologies:

• AI
• AWS
• EC2
• Firmware
• Hardware
• Machine Learning
• Cloud
• Support

More

We are Annapurna Labs, part of AWS, and we develop hardware and software components for EC2 infrastructure. Our team builds networking solutions for machine learning and high-performance computing workloads on AWS. We work in a mixed-discipline, international environment alongside infrastructure experts, hardware engineers, RTL engineers, scientists, and architects. We value mentorship, fast-paced innovation, work-life balance, and continuous learning, with opportunities to grow alongside principal-level engineers and directors. The role offers the chance to work on cutting-edge AI training hardware and contribute to one of Amazons most important initiatives. The position is based in Cupertino, California, with a listed annual base salary range of $165,600, and our package includes sign-on payments, restricted stock units, and comprehensive benefits such as health coverage, 401(k) matching, paid time off, and parental leave.

last updated 43 week of 2026