LLVM/MLIR Compiler Engineer for AI Accelerators

Meta; Bellevue, Menlo Park, New York (WA, CA, NY); On-site; USD 180,000 - 240,000; Full time

Job summary

Meta in Bellevue, WA seeks a Software Engineer to join the MTIA LLVM Compiler team, building a production compiler stack for Meta's AI accelerators. You will work across compiler, ML, firmware, and ASIC experts to optimize deep learning model compilation and generate efficient binaries for hardware blocks.

The role requires 5+ years in compiler development, proficiency in C++ or Rust, and strong leadership to drive cross-team initiatives.

Qualifications

• Bachelor's degree or equivalent; 5+ years in compiler or code optimization.
• Proficient in C++ or Rust with large-scale software development experience.
• Experience with LLVM/MLIR or similar production compiler codebase.
• Track record delivering major compiler features or optimization passes.
• Experience driving cross-team technical initiatives and roadmap influence.
• Ability to mentor engineers and raise technical standards.
• Experience with hardware architectures (SIMD, GPUs, AI accelerators).
• Familiarity with ML frameworks (PyTorch, TensorFlow).

Responsibilities

• Architect, design, and implement production compiler stack for AI accelerators.
• Collaborate across compiler, ML, firmware, and ASIC teams.
• Optimize code generation and back-end compiler passes for hardware blocks.
• Contribute to upstream LLVM/MLIR projects.
• Mentor engineers and influence roadmap decisions.

Skills

C++

Rust

Compiler development

Performance analysis

Debugging

Code optimization

LLVM/MLIR experience

Cross-team collaboration

Mentoring

AI frameworks experience

Hardware acceleration knowledge

Education

Bachelor's degree in Computer Science/Engineering or equivalent

Tools

LLVM

MLIR

GCC

MSVC

Job description

Meta in Bellevue, WA seeks a Software Engineer to join the MTIA LLVM Compiler team, building a production compiler stack for Meta's AI accelerators. You will work across compiler, ML, firmware, and ASIC experts to optimize deep learning model compilation and generate efficient binaries for hardware blocks.

The role requires 5+ years in compiler development, proficiency in C++ or Rust, and strong leadership to drive cross-team initiatives.