Machine Learning - Compiler Engineer , AWS Neuron, Annapurna Labs
Core
Building next-generation compilers to optimize ML models (LLMs, vision transformers) for deployment on AWS Inferentia and Trainium custom chips.
Role type
Senior IC compiler engineer (ML accelerators)
Builds
High-performance Neuron compiler and SDK for AWS accelerators
Domain
Cloud infrastructure + ML compilers + Custom silicon
Deliverable
production ML models
Required skills
C++, compiler design, system architecture, ML framework internals, numerical error analysis
Preferred skills
OpenXLA, StableHLO, MLIR, LLVM, PyTorch, JAX, TVM, Bazel/CMake
Technologies
AWS Inferentia, AWS Trainium, PyTorch, TensorFlow, JAX, StableHLO, OpenXLA, MLIR, LLVM, Bazel, CMake
Responsibilities
Design and implement compiler optimization and verification passes, build features to surface accelerator peculiarities to developers, implement tools to analyze numerical errors, resolve root causes of compiler defects, partner with chip architects and ML teams
Seniority
Senior, hands-on IC
