Machine Learning Compiler Engineer, Annapurna Labs

The AWS Neuron Compiler team is seeking talented compiler engineers to develop a cutting-edge deep learning compiler stack. This role focuses on optimizing application models across various domains, including Large Language Models and Vision, working with frameworks like PyTorch, TensorFlow, and JAX.

As a Machine Learning Compiler Engineer, you'll work with AWS's custom-built ML accelerators (Inferentia/Trainium) that power advanced ML capabilities including Generative AI. You'll be responsible for designing, developing, and optimizing compiler features while collaborating with cross-functional teams.

The position involves working on critical aspects like instruction scheduling, memory allocation, data transfer optimization, graph partitioning, parallel programming, and code generation. You'll also participate in hardware-software co-design and new hardware bring-up initiatives.

The role offers the opportunity to solve unique technical challenges in a startup-like environment while impacting AWS's global customer base. You'll work alongside talented engineers, participate in design discussions, conduct code reviews, and help drive business decisions through technical expertise.

This position at Annapurna Labs, an AWS subsidiary, offers competitive compensation ranging from $129,300 to $223,600 based on location and experience, plus additional benefits including medical insurance and 401k. The role is based in Cupertino, CA, and requires at least 3 years of professional software development experience with strong expertise in C++ or Python.