Software Engineer II - AI/ML, Neuron Inference

The Annapurna Labs team at Amazon Web Services (AWS) builds the AWS Neuron SDK, the software development kit used to accelerate deep learning and GenAI workloads on AWS Trainium machine learning accelerators. The Neuron Inference organization is at the forefront of optimizing inference performance for a wide range of ML models and model architectures on AWS's custom ML accelerators. We are working across the stack from PyTorch to the hardware-software boundary, our engineers build systematic infrastructure, develop high-performance kernels for ML functions, ensuring every compute unit is fine-tuned for optimal performance forour customers' demanding workloads. We combine deep hardware knowledge with ML expertise to push the boundaries of what's possible in AI acceleration. As part of the broader Neuron organization, our team works across multiple technology layers — from frameworks and kernels through to compiler, runtime, and collectives. We not only optimize performance on open weight models but also contribute to future architecture designs and work closely with customers to enable optimal performance on their models. This role offers a unique opportunity to work at the intersection of machine learning, high-performance computing, and distributed architectures, where you'll help shape the future of AI acceleration technology. As a Software Engineer II - AI/ML, you will design and implement business-critical features and own the performance of models on AWS Trainium end to end. We operate in spaces that are very large, yet our teams remain small and agile. There is no blueprint. We're inventing. We're experimenting. It is a very unique learning culture. The team works closely with customers on their model enablement, providing direct support and optimization expertise to ensure their machine learning workloads achieve optimal performance on AWS ML accelerators. The team also collaborates with open source ecosystems to provide seamless integration and bring peak performance at scale for customers and developers. This role is responsible for development, enablement, and performance tuning of a wide variety of LLM model families, including massive-scale large language models like the GLM, GPT-OSS, Kimi, and beyond. The Model Enablement team works side by side with compiler engineers and runtime engineers to create, build, and tune distributed inference solutions with AWS Trainium. You can learn more about Neuron: * https://awsdocs-neuron.readthedocs-hosted.com/en/latest/neuron-guide/neuron-cc/index.html * https://aws.amazon.com/machine-learning/neuron/ * https://github.com/aws/aws-neuron-sdk * https://www.amazon.science/how-silicon-innovation-became-the-secret-sauce-behind-awss-success Key job responsibilities This role contributes to building distributed inference support for PyTorch in the Neuron SDK, and tunes models to ensure the highest performance and maximize efficiency running on AWS Trainium. Strong software development in Python, system-level programming, and ML knowledge are all critical to this role. Our engineers collaborate across compiler, runtime, framework, and hardware teams to optimize machine learning workloads for our global customer base. Working at the intersection of software, hardware, and machine learning systems, you'll bring expertise in low-level optimization, system architecture, and ML model acceleration. In this role, you will: * Design, develop, and optimize machine learning models and frameworks for deployment on custom ML hardware accelerators. * Participate in all stages of the ML system development lifecycle, including distributed-computing-based architecture design, implementation, performance profiling, hardware-specific optimizations, testing, and production deployment. * Build infrastructure to systematically analyze and onboard multiple models with diverse architectures. * Understand and produce NKI (Neuron Kernel Interface) kernels, and design and implement high-performance kernels and features for ML operations, leveraging the Neuron architecture and programming models. * Analyze and optimize system-level performance across multiple generations of Neuron hardware, using knowledge of the underlying Trainium hardware architecture to guide optimization decisions. * Conduct detailed performance analysis using profiling tools to identify and resolve bottlenecks. * Implement optimizations such as fusion, sharding, tiling, and scheduling. * Conduct comprehensive testing, including unit and end-to-end model testing with continuous deployment and releases through pipelines. * Work directly with customers to enable and optimize their ML models on AWS accelerators. * Collaborate across teams to develop innovative optimization techniques, and share proposals, findings, and learnings with internal developers and external customers via high-quality designs and documentation. A day in the life You will collaborate with a cross-functional team of applied scientists, systems engineers, and product managers to deliver state-of-the-art inference capabilities for Generative AI applications. Your work will involve debugging performance issues, optimizing memory usage, and contributing to the future of Neuron's inference stack across Amazon and the Open Source Community. As you design and code solutions to help our team drive efficiencies in software architecture, you'll create metrics, implement automation and other improvements, and resolve the root cause of software defects. You will also build high-impact solutions to deliver to our large customer base and participate in design discussions and code reviews, and communicate with internal and external stakeholders. You will work cross-functionally to help drive business decisions with your technical input. You will work in a startup-like development environment, where you're always working on the most important initiative. About the team The Model Enablement team fosters a builder's culture where experimentation is encouraged and impact is measurable. We emphasize collaboration, technical ownership, and continuous learning. Our team is dedicated to supporting new members. We have a broad mix of experience levels and tenures, and we're building an environment that celebrates knowledge sharing and mentorship. We care about your career growth and strive to assign projects that help our team members develop their engineering expertise so they feel empowered to take on more complex tasks in the future. Join us to solve some of the most interesting and impactful infrastructure challenges in AI/ML today. - 3+ years of non-internship professional software development experience - 3+ years of non-internship design or architecture (design patterns, reliability and scaling) of new and existing systems experience - 1+ years of designing and developing large-scale, multi-tiered, multi-threaded, embedded or distributed software applications, tools, systems, and services using: C#, C++, Java, or Perl experience - Bachelor's degree or foreign equivalent in Computer Science, Engineering, Mathematics, or a related field - Experience debugging, profiling, and implementing best software engineering practices in large-scale systems - Experience with Machine Learning and Large Language Model fundamentals, including architecture, training/inference lifecycles, and optimization of model execution - Knowledge of Python and/or C++ programming - Strong understanding of system performance, memory management, and parallel computing principles. - Experience with CUDA kernels or ML/low-level kernels, or experience in performant kernel development (CUTLASS, FlashInfer) - Experience with vLLM, SGLang, TensorRT or similar platforms in production environments - Knowledge of syntax and tile-level semantics similar to Triton - Knowledge of computer architecture, operating systems, and parallel computing - Experience with PyTorch, JIT compilation, and AOT tracing - Knowledge of machine learning model architecture