Engineering Manager, ML Compilers (Triton / Kernel DSLs)

Job Title

Engineering Manager, ML Compilers (Triton / Kernel DSLs)

Role Summary

Lead a team building Meta's ML compiler stack for custom AI accelerators, focusing on Triton and lower-level domain-specific languages that enable efficient kernel execution on MTIA hardware.

The role spans compiler infrastructure, kernel libraries, production deployment across accelerator generations, open-source contributions, and close partnership with hardware, runtime, framework, and model teams.

Experience Level

Senior - requires leadership experience. Minimum: 5+ years managing a software team and experience with compiler architecture and ML compilers or DSLs.

Responsibilities

Drive technical delivery and team leadership across compiler and kernel toolchains.

• Own development of the DSL compiler toolchain and ML compiler integration for MTIA accelerators.
• Enable GEMM and non-GEMM kernels, collectives support, and low-level DSL targets.
• Lead multiple engineering teams and technical leads to deliver large-scale projects across compiler infra, kernel libraries, and deployments.
• Remain technically engaged and contribute hands-on to guide technical direction and quality.
• Collaborate with hardware/architecture, runtime, framework, kernel, and model teams to shape strategy and roadmap.
• Balance short-term deployment commitments with long-term compiler investments and influence ISA/architecture decisions.
• Mentor engineers, uphold high engineering craft standards, and represent work externally (open source, conferences, publications).

Requirements

Must-have technical and leadership qualifications, followed by common preferred skills.

• Must-have: 5+ years managing a software team in a fast-paced environment.
• Must-have: Experience with compiler architecture and development, especially ML compilers or domain-specific languages (DSLs).
• Must-have: Proven experience delivering full product lifecycles: prototyping, deployment, and support.
• Preferred: Experience with Triton, MLIR, or LLVM and contributing upstream to open-source projects.
• Preferred: Familiarity with hardware architectures (SIMD, GPU, RISC-V, AI accelerators) and influencing architecture roadmaps.
• Preferred: Knowledge of PyTorch compiler internals (e.g., Inductor) and GPU programming models (CUDA, OpenCL, OpenMP).
• Preferred: Kernel-level performance optimization for AI accelerators and experience integrating AI tools to improve workflows.
• Preferred: Experience implementing responsible AI practices (risk assessment, bias mitigation, quality reviews).

Education Requirements

Not specified.

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.