Software Engineer, LLVM Compiler

Job Title

Software Engineer, LLVM Compiler

Role Summary

Work on the MTIA LLVM Compiler team to architect, design, and implement a production compiler stack targeting Meta's custom silicon AI accelerators. The role spans compiling PyTorch models through LLVM/MLIR down to optimized binaries for hardware accelerator blocks.

Collaborate with compiler engineers, ML framework teams, firmware, and hardware architects to deliver correct, high-performance code generation and tooling for training and inference workloads.

Experience Level

Mid-level - the role expects experienced engineers (documented minimum: 5+ years in compiler or code optimization development).

Responsibilities

Key responsibilities include:

• Design and implement LLVM compiler passes: optimization, lowering, and code generation for ML accelerator backends.
• Analyze and improve compiler-generated code quality via profiling, benchmarking, and performance instrumentation across ML workloads.
• Develop and maintain LLVM IR/MLIR transformations to map ML computation graphs to hardware execution units efficiently.
• Collaborate with hardware architecture and ML framework teams to define compiler interfaces and ensure correctness on custom silicon.
• Diagnose and fix correctness and performance regressions through systematic debugging and root cause analysis.
• Own technical design trade-offs between compilation time, runtime performance, and code size for inference and training.
• Build and improve automated testing and benchmarking infrastructure to validate compiler correctness and track performance across hardware generations.
• Drive compiler best practices, contribute to documentation, and mentor other engineers on LLVM and ML-specific code generation.

Requirements

Must-have skills and experience:

• 5+ years developing compilers or code optimization software with demonstrated technical leadership.
• Proficiency in C++ or Rust and experience in large-scale software development, debugging, testing, and performance analysis.
• Experience with LLVM/MLIR or a comparable production compiler codebase (e.g., GCC, MSVC).
• Proven track record delivering significant compiler features or optimization passes end-to-end.
• Experience driving cross-team technical initiatives and working across hardware, ML frameworks, and runtime systems.
• Experience accelerating deep learning models on hardware architectures and mentoring engineers.

Nice-to-have:

• Experience with upstream LLVM/MLIR contributions, deep learning model compilation, operator fusion, tiling, quantization-aware compilation, or hardware-specific optimizations.
• Familiarity with SIMD, GPU, RISC-V, AI accelerators, or mainstream ML frameworks such as PyTorch or TensorFlow.
• Experience with machine-code generation/back-end optimizations (instruction selection, register allocation, scheduling) and integrating AI tools to improve workflows.
• Experience implementing responsible AI practices (risk assessment, bias mitigation, quality reviews).

Education Requirements

Bachelor's degree in Computer Science, Computer Engineering, or a relevant technical field, or equivalent practical experience.

About the Company

Company: Meta Platforms

Headquarters: Menlo Park, California, United States

American technology company that develops social networking products (Facebook, Instagram, WhatsApp) and invests in virtual/augmented reality hardware and software through Reality Labs, focusing on connectivity, advertising, and immersive computing experiences.