Staff Mojo Compiler Engineer

Modular•United States / Canada,

117d•$216,000 - $372,000•Hybrid

About The Position

This is a technical leadership (IC) role in collaboration with the compiler engineering manager. You will be driving the technical roadmap for the Mojo compiler backend, creating new features, proposing new ideas, driving code reviews, and leading design documents. Mentoring other engineers is a key element of this role. Although your primary focus is the compiler backend code generation and lowering across multiple devices and architectures, you will also grow the Mojo tooling capabilities, in particular the debugger. This is where your mentoring and technical leading skills will shine: You will be driving junior engineers working in tooling, training them and helping them grow and lifting the team's overall capability there. Candidates based in the US, Canada, and UK are welcome to apply. You can work in our office in Los Altos, CA, or Edinburgh, or remotely from home. Onboarding for new hires is conducted in-person at the appropriate office.

Requirements

Deep, hands-on compiler experience; IR design, optimization passes, and code generation in a production toolchain.
Strong experience with both LLVM and MLIR — comfortable across dialects, lowering pipelines, and codegen.
A track record of technical leadership without formal management: leading projects, setting architecture, and growing other engineers.
Strong C++ and Python.
Hands-on experience with debuggers in the LLDB and GDB family; not just using them, but ideally extending or building debugger support for a language or runtime.
Familiarity with editor/IDE tooling, language servers (LSP), VS Code extensions, or comparable developer-experience surfaces.
Working knowledge of CPU and/or GPU/accelerator architecture and the perf-analysis mindset that goes with it.
Excellent collaboration and communication across distributed, cross-functional teams.

Nice To Haves

Experience working in open source projects, user-facing SDK or community driven projects, actively involved in user interactions and requirement capture
Solid command of profilers and performance analysis (e.g. Linux perf, LLVM XRay, or comparable sampling/instrumentation tooling).
Accelerator backend code generation — experience targeting GPUs or custom AI accelerators at the codegen level.
Graph compiler experience — building or optimizing graph-level compilation (e.g. XLA, TVM, or comparable ML compiler stacks).
GPU debugging — hands-on with cuda-gdb and/or ROCgdb, readable symbol mangling and backtraces, and visualizing GPU data types (distinct from GPU codegen).
Contributions to upstream LLVM/MLIR or other open-source compiler communities.

Responsibilities

Own the technical strategy and architecture for the Mojo compiler backend: IR lowering, optimization passes, and code generation targeting heterogeneous hardware. In Mojo, this spans from MLIR all the way to LLVM backend or custom compilers.
Act as tech lead for a team of backend compiler engineers, breaking down ambiguous problems, driving design reviews, setting technical standards, and multiplying the team's output.
Drive performance: use and extend profiling tools to find optimization opportunities, then design the compiler passes that capture them.
Lead the team's developer-tooling effort, debuggers (LLDB and the GDB family), profilers, debug-info generation, and data introspection / visualizers — and bring other engineers up to speed as contributors and owners in that space.
Steer the editor and language-tooling experience in partnership with the rest of the compiler team, where the boundary between the language server (LSP) and the IDE extension is deliberately blurry.
Partner across the Mojo language, MAX, kernel, and hardware-enablement teams, and engage Modular's open-source community.
Mentor engineers, raise the technical bar, and influence direction beyond your immediate team.