2026 ML Compiler Software Engineer PEY Co-op (12-16 months), AWS Neuron, Annapurna Labs

Amazon Development Centre Canada ULC - K03; Toronto; On-site; CAD 100,000 - 125,000; Full time

Job summary

A leading technology company in Toronto is seeking a student intern to contribute to ML workload optimization. This role includes building key features, analyzing performance, and optimizing compilers. Ideal candidates are pursuing a related degree with an expected graduation between May 2027 and May 2028. The internship offers mentorship from experienced engineers and a chance to work on impactful technology.

Qualifications

• Enrolled in an academic program physically located in Canada.
• Anticipated graduation date between May 2027 - May 2028.
• Strong interests in code generation and ML frameworks.

Responsibilities

• Contribute to making Neuron best-in-class for ML workloads.
• Build business-critical features and analyze performance.
• Develop compiler or kernel optimizations.

Skills

Knowledge of code generation

Compiler optimization

Experience with PyTorch

Kernel development

Education

Bachelor’s degree or higher in Computer Science or related fields

Job description

Overview

At AWS, our mission is to make deep learning accessible to every developer by democratizing access to cutting-edge infrastructure. To achieve this, we’ve built custom silicon (AWS Inferentia and Trainium) and the AWS Neuron SDK that together deliver high-performance, cost-effective machine learning in the cloud.

The AWS Neuron SDK includes a compiler, runtime, debugger, and libraries integrated with popular frameworks such as PyTorch and TensorFlow. It is preinstalled in AWS Deep Learning AMIs and Containers so customers can quickly get started with training and inference on AWS ML accelerators.

The Neuron Toronto team focuses on performance, kernels, and tooling—analyzing and optimizing end-to-end ML workloads, developing and maintaining highly optimized kernels, and building performance modeling, profiling, and accuracy debugging tools. Together, these efforts ensure that Neuron delivers best-in-class performance, flexibility, and usability for customers deploying large-scale machine learning models.

Responsibilities

As a student intern, you will contribute to the efforts that make Neuron best-in-class for ML workloads. You’ll gain hands-on experience building business-critical features, analyzing performance, developing compiler or kernel optimizations, and building tools that provide deep insights into model execution. You’ll be mentored by experienced engineers while working on technology that directly accelerates customer workloads at scale.

BASIC QUALIFICATIONS

• Are enrolled in a academic program that is physically located in Canada
• Are enrolled in a Bachelor’s degree or higher in Computer Science, Engineering Science, Computer Engineering, Electrical Engineering, or majors relating to these fields with an anticipated graduation date between May 2027 - May 2028
• Strong interests and academic qualifications / research focus in two of the following: 1) Knowledge of code generation, compute graph optimization, resource scheduling 2) Compiler - Optimizing compilers (internals of LLVM, clang, etc) 3) Machine Learning frameworks (PyTorch, JAX) and Machine Learning (Experience with XLA, TVM, MLIR, LLVM) 4) Kernel development—experience writing CUDA kernels, OpenCL kernels, or ML-specific kernels
• Available for a 12-16-month internship starting May 2026

PREFERRED QUALIFICATIONS

• Previous technical internship(s) related to the areas of interest / research focus listed above
• Experience in optimization mathematics such as linear programming and nonlinear optimization
• Academic coursework in Compiler Design / Construction, Programming Language Theory, Computer Architecture, Advanced Algorithms & Data Structures

Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

Our inclusive culture empowers Amazonians to deliver the best results for our customers.