ML Compiler Engineer II - Neuron Kernel Interface , Annapurna Labs

ML Compiler Engineer II role on the AWS Neuron Compiler team to develop and maintain tooling that improves compiler accuracy and reliability. The engineer will build fuzzers and specification synthesis tooling for an LLVM-based compiler and collaborate with a science-focused team delivering for customers working on Generative AI.

• Develop and maintain Neuron Compiler tooling for accuracy and reliability
• Build fuzzers and specification synthesis tooling for an LLVM-based compiler
• Collaborate with a science-focused team to improve compiler capabilities
• Use strong C++/Python software development to implement compiler tooling
• Support optimization of deep learning models for Inferentia/Trainium accelerators

Work on the AWS Neuron Compiler deep learning compiler stack that optimizes models from PyTorch, TensorFlow, and JAX for AWS custom accelerators like Inferentia and Trainium. Strong C++/Python development is required, with compiler tooling expertise centered on LLVM-based infrastructure, fuzz testing, and specification synthesis to validate correctness and improve reliability.

The ideal candidate is an ML Compiler Engineer II with 3+ years of professional software development experience and 2+ years designing or architecting systems with an emphasis on design patterns, reliability, and scaling. They are strong in C++/Python and have experience working with compiler tooling (LLVM-based), including fuzzers and specification synthesis; a background in compiler development or ML/AI accelerators is preferred.