Principal Compiler Architect for MTIA Software

Meta; Menlo Park (CA); On-site; USD 219,000 - 301,000; Full time

Job summary

Meta is seeking a principal-level Compiler Architect to lead MTIA compiler infrastructure across the full stack, from graph transformations to code generation. You will tackle ML workload analysis, memory hierarchy management, and performance optimizations on Meta's custom AI accelerators.

The role blends research with production-quality engineering and cross-team collaboration. The ideal candidate has deep accelerator experience, leadership in multi-year projects, and a strong track record in

Qualifications

• Bachelor's degree in Computer Science, Computer Engineering, or related field or equivalent practical experience.
• Experience in compiler development, code generation, or accelerator optimization.
• Experience architecting production compiler infrastructure for ML accelerators or custom silicon.
• Knowledge of compiler IRs, optimization passes, and code generation techniques.
• Experience leading multi-year cross-functional technical initiatives.

Responsibilities

• Design compiler architecture across the MTIA compiler stack, including graph IRs and code generation.
• Build extensible compiler frameworks and representations for ML model evolution.
• Eliminate performance bottlenecks through graph optimizations and scheduling.
• Collaborate with MTIA hardware teams on hardware-software co-design improvements.
• Mentor engineers and lead architecture reviews across teams.
• Integrate MTIA compiler with ML frameworks like PyTorch.
• Apply state-of-the-art compiler technologies such as MLIR and XLA.

Skills

Compiler design

Performance optimization

Hardware-software co-design

MLIR

Code generation

Cross-functional leadership

Education

Bachelor's degree in CS/CE

Master's or PhD in CS/CE

Tools

MLIR

XLA

TVM

Glow

Job description

Meta is seeking a principal-level Compiler Architect to lead MTIA compiler infrastructure across the full stack, from graph transformations to code generation. You will tackle ML workload analysis, memory hierarchy management, and performance optimizations on Meta's custom AI accelerators.

The role blends research with production-quality engineering and cross-team collaboration. The ideal candidate has deep accelerator experience, leadership in multi-year projects, and a strong track record in