Software Dev Engineer, ECNitro; AWS

Job  | Amazon Development Center U.S., Inc.

Join the EC2 Nitro Machine Learning Systems team to revolutionize accelerated computing in the cloud. We’re seeking an exceptional Software Development Engineer to build and optimize the performance measurement infrastructure for some of the most computationally intensive AI/ML workloads on AWS. In this role, you’ll establish EC2 as the definitive source for best‑known‑configurations across diverse ML applications including LLMs, multimodal models, and video generation workloads.

Your expertise will directly influence future platform designs by translating performance insights from state‑of‑the‑art research and customer workloads into technical requirements for upcoming accelerated platform launches.

Your impact will extend from low‑level systems (CUDA, EFA, firmware) through ML frameworks to serving layers, requiring deep technical knowledge and the ability to communicate complex performance data as actionable business insights. This position offers the unique opportunity to shape the future of machine learning infrastructure at cloud scale while working at the intersection of high‑performance computing, distributed systems, and machine learning technologies.

• Design and build foundational infrastructure for ML performance measurement that scales with business demand and operates as reliable CI/CD systems, ensuring high‑quality implementations that balance customer requirements with operational excellence.
• Develop comprehensive regression test coverage across all major component releases—including frameworks, firmware, drivers, and networking technologies—to maintain optimal platform performance.
• Collaborate with cross‑functional teams to establish EC2 as the definitive source for best‑known‑configurations across diverse ML applications including LLMs, multimodal models, and MoE architectures.
• Document and communicate performance insights to influence future platform designs by translating technical findings from research and customer workloads into actionable recommendations.
• Identify and resolve complex performance challenges through systematic analysis of training and inference performance KPIs across accelerated platforms, working directly with customers to improve their ML system efficiency.

Your typical day begins with reviewing performance data from overnight benchmark runs across various ML frameworks and hardware configurations. You’ll investigate anomalies, collaborate with the team on optimization opportunities, and join design reviews to influence future platform capabilities. You’ll balance your time between building measurement infrastructure, analyzing performance trends, and documenting best practices to help customers optimize their workloads.

The EC2 Nitro Machine Learning Systems team is responsible for development, operations, and maintenance of scale‑out machine learning platforms used for training and inference workloads. We build and optimize the infrastructure that powers some of the most computationally intensive AI/ML workloads in the cloud. Our team is passionate about creating reliable, high‑performance systems that enable customers to push the boundaries of what’s possible with machine learning.

• 3+ years of non‑internship professional software development experience.
• 2+ years of non‑internship design or architecture (design patterns, reliability and scaling) of new and existing systems experience.
• Experience programming with at least one software programming language.
• Knowledge of Machine Learning and LLM fundamentals, including transformer architecture, training/inference life cycles, and optimization techniques.

• 3+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience.
• Bachelor’s degree in computer science or equivalent.
• Knowledge of ML frameworks including JAX, PyTorch, vLLM, SGLang, Dynamo, Torch
  
  XLA, and Tensor
  
  RT.
• Knowledge of machine learning model architecture and inference.

Amazon is an equal opportunity employer and does not…