ML Compiler Engineer for Custom AI Accelerators

Meta; Sunnyvale (CA); On-site; USD 184,000 - 257,000; Full time

Job summary

Reality Labs (RL) at Meta is hiring a Software Engineer, Systems ML - Compilers in Sunnyvale to lead architecture and implementation of ML compiler infrastructure for custom AI accelerators. You will drive IR design, optimization passes, and code generation for PyTorch models targeting silicon, collaborating with hardware teams for edge devices.

The role emphasizes LLVM/MLIR-based toolchains, end-to-end delivery, and cross‑team coordination, with a focus on performance, power efficiency, and

Qualifications

• Bachelor's degree in Computer Science, Computer Engineering, or related field.
• 3+ years of experience in developing compilers, toolchains, or code optimization software.
• Experience with LLVM, MLIR, or similar compiler infrastructure frameworks.
• Experience in designing intermediate representations and implementing compiler optimization passes.
• Experience with hardware architectures such as GPUs, TPUs, or custom AI accelerators.
• Experience with software development using C++ for compiler and systems‑level programming.
• Experience leading end‑to‑end technical design and delivery of compiler infrastructure initiatives across multiple teams.

Responsibilities

• Lead the architecture and implementation of ML compiler infrastructure, including intermediate representations (IR), optimization passes, and code generation targeting custom AI accelerators.
• Design and implement compiler transformations informed by hardware architecture constraints for GPU, TPU, and edge AI accelerators.
• Drive the development of LLVM/MLIR-based toolchains for compiling PyTorch models to optimized binaries for custom silicon.
• Work with hardware architects to co‑design compiler features that maximize performance, power efficiency, and programmability for edge devices.
• Analyze and improve the efficiency, scalability, and stability of compiler toolchains, ensuring they can be extended to new hardware targets.
• Lead technical roadmapping for compiler infrastructure initiatives, coordinate execution across teams, and mentor engineers on compiler design patterns.
• Conduct design and code reviews, evaluate code performance, and drive resolution of compiler and cross‑disciplinary system issues.
• Interface with other compiler‑focused teams (PyTorch, ExecuTorch) to evaluate and incorporate innovations.

Skills

LLVM/MLIR experience

C++ for systems programming

Compiler infrastructure design

End-to-end compiler delivery

PyTorch familiarity

Education

Bachelor's degree in Computer Science or related field

Tools

LLVM

MLIR

Job description

Reality Labs (RL) at Meta is hiring a Software Engineer, Systems ML - Compilers in Sunnyvale to lead architecture and implementation of ML compiler infrastructure for custom AI accelerators. You will drive IR design, optimization passes, and code generation for PyTorch models targeting silicon, collaborating with hardware teams for edge devices.

The role emphasizes LLVM/MLIR-based toolchains, end-to-end delivery, and cross‑team coordination, with a focus on performance, power efficiency, and