Software Engineer, Systems ML - Frameworks / Compilers / Kernels

Meta is seeking a Software Engineer for their MTIA (Meta Training & Inference Accelerator) Software team, which is part of the industry-leading PyTorch AI framework organization. This role offers an exciting opportunity to work on cutting-edge AI infrastructure and compiler development.

The position involves developing crucial components of Meta's AI infrastructure, including PyTorch framework components, AI compiler and runtime systems, and high-performance kernels. You'll be working directly with AI researchers to analyze and optimize deep learning models for MTIA hardware platforms, while also collaborating with hardware design teams to develop efficient compiler optimizations.

As a member of this team, you'll be responsible for implementing software development best practices, designing features, and performing optimization and performance tuning. The role requires expertise in C/C++ programming and experience with AI framework development or model acceleration on hardware architectures.

The ideal candidate should have at least 4 years of experience or equivalent advanced degrees in Computer Science or related fields. Knowledge of GPU, CPU, or AI hardware accelerator architectures is preferred, along with experience in frameworks like PyTorch, TensorFlow, or ONNX. Specialized knowledge in AI high-performance kernels (CUDA, OpenMP) or AI compiler optimization (MLIR, LLVM) is highly valued.

This position offers competitive compensation ranging from $56.25/hour to $173,000/year, plus bonus, equity, and benefits. You'll be working at Meta's Bellevue location, contributing to the company's mission of advancing AI technology and social connection platforms. This is an excellent opportunity for someone passionate about AI infrastructure and interested in shaping the future of machine learning systems at scale.