ML Systems Software Development Engineer Intern, Annapurna Labs - 2027

About the team Annapurna Labs designs the custom silicon at Amazon - including Graviton (our server processors), Trainium and Inferentia (our machine learning training and inference accelerators), and the Nitro system that powers modern EC2. We own the full stack, from silicon through the software that makes it work (compilers, runtimes, drivers, and the frameworks that let customers run workloads on our hardware). As a Software Development Engineer Intern, you'll write the software that turns custom silicon into products used by millions. Basic qualifications - Currently enrolled in a Bachelor's degree program or higher in Computer Science, Computer Engineering, Electrical Engineering, or a related field - Programming experience through coursework, research, or a previous internship using Python, C, and/or C++ - Strong interest and academic, research, or project experience in at least two of the following areas: 1/ Performance engineering, profiling, benchmarking, or low-level systems optimization. 2/ Kernel development, parallel programming, or computer architecture. 3/ Developer tooling, including profilers, debuggers, diagnostics, or visualization. 4/ Data structures and algorithms. 5/ Machine learning frameworks and models, including PyTorch or JAX. 6/ Compiler or ML systems technologies such as LLVM, MLIR, XLA, or TVM Preferred qualifications - Experience bringing up and optimizing machine learning models - Experience writing or optimizing kernels for GPUs, ML accelerators, or FPGAs - Experience using performance analysis tools or building developer tools - Previous technical internship or relevant research experience - Experience with full stack development, including front-end technologies such as TypeScript and React, and back-end services (experience with Go is a plus) - Ability to explain technical challenges and solutions clearly - Ability to independently work through ambiguous or undefined problems and think abstractly - Experience in developing agentic workflows for work automation

- Currently enrolled in a Bachelor's degree program or higher in Computer Science, Computer Engineering, Electrical Engineering, or a related field - Programming experience through coursework, research, or a previous internship using Python, C, and/or C++ - Strong interest and academic, research, or project experience in at least two of the following areas: 1/ Performance engineering, profiling, benchmarking, or low-level systems optimization. 2/ Kernel development, parallel programming, or computer architecture. 3/ Developer tooling, including profilers, debuggers, diagnostics, or visualization. 4/ Data structures and algorithms. 5/ Machine learning frameworks and models, including PyTorch or JAX. 6/ Compiler or ML systems technologies such as LLVM, MLIR, XLA, or TVM Preferred qualifications - Experience bringing up and optimizing machine learning models - Experience writing or optimizing kernels for GPUs, ML accelerators, or FPGAs - Experience using performance analysis tools or building developer tools - Previous technical internship or relevant research experience - Experience with full stack development, including front-end technologies such as TypeScript and React, and back-end services (experience with Go is a plus) - Ability to explain technical challenges and solutions clearly - Ability to independently work through ambiguous or undefined problems and think abstractly - Experience in developing agentic workflows for work automation Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

What's in it for you? An internship at Amazon means real responsibility from day one. You'll build, test, and learn alongside people who want to see you succeed, while making an impact that reaches far beyond campus. Key job responsibilities Join the Toronto Neuron team (Annapurna Labs) to bring up and optimize state-of-the-art machine learning models for peak performance on AWS Trainium and build industry-leading profiling tools that help engineers identify and fix bottlenecks. The Toronto Neuron team is hiring 2027 interns in two focus areas: - Frontier Model Performance team. Take newly released machine learning models from first bring-up to peak performance on current and next-generation AWS Trainium silicon. You could write and tune kernels, optimize sharding and model execution, build benchmark and measurement infrastructure, or tackle architectural bottlenecks that shape future Trainium designs. The best ideas do not stay trapped in one model. The team turns them into reusable Neuron components and optimization techniques for flagship open-source and customer workloads.

Key job responsibilities Join the Toronto Neuron team (Annapurna Labs) to bring up and optimize state-of-the-art machine learning models for peak performance on AWS Trainium and build industry-leading profiling tools that help engineers identify and fix bottlenecks. The Toronto Neuron team is hiring 2027 interns in two focus areas: - Frontier Model Performance team. Take newly released machine learning models from first bring-up to peak performance on current and next-generation AWS Trainium silicon. You could write and tune kernels, optimize sharding and model execution, build benchmark and measurement infrastructure, or tackle architectural bottlenecks that shape future Trainium designs. The best ideas do not stay trapped in one model. The team turns them into reusable Neuron components and optimization techniques for flagship open-source and customer workloads. - Developer Experience (DevEx) team. Make AWS Trainium performance visible. The team owns Neuron Explorer and builds state-of-the-art profiling, debugging, and analysis tools that let engineers see how efficiently models and kernels use the hardware, pinpoint bottlenecks, and decide what to optimize next. You could build low-level performance data collection, analysis engines, interactive visualizations, or developer workflows. The goal is to cut through complex execution data so customers can debug faster and reach higher performance on Trainium. Interns will join one of these teams based on their interests and experience. In either role, you will solve real customer and engineering problems with mentorship from experienced machine learning and systems engineers.

The Toronto Neuron team is hiring 2027 interns in two focus areas: - Frontier Model Performance team. Take newly released machine learning models from first bring-up to peak performance on current and next-generation AWS Trainium silicon. You could write and tune kernels, optimize sharding and model execution, build benchmark and measurement infrastructure, or tackle architectural bottlenecks that shape future Trainium designs. The best ideas do not stay trapped in one model. The team turns them into reusable Neuron components and optimization techniques for flagship open-source and customer workloads. - Developer Experience (DevEx) team. Make AWS Trainium performance visible. The team owns Neuron Explorer and builds state-of-the-art profiling, debugging, and analysis tools that let engineers see how efficiently models and kernels use the hardware, pinpoint bottlenecks, and decide what to optimize next. You could build low-level performance data collection, analysis engines, interactive visualizations, or developer workflows. The goal is to cut through complex execution data so customers can debug faster and reach higher performance on Trainium. Interns will join one of these teams based on their interests and experience. In either role, you will solve real customer and engineering problems with mentorship from experienced machine learning and systems engineers. Internship Options - 12-16 month internship (starts May 2027) - 3-4 month internship (starts January 2027, May 2027, September 2027)