Principal GPU Compiler Engineer

Help define the compiler stack behind next-generation AI GPU technology.You'll join a well-funded company developing programmable GPU IP for AI workloads. As the software platform matures towards production, they're looking for a Principal GPU Compiler Engineer to help shape the compiler architecture that bridges modern ML frameworks with custom GPU hardware.What you'll doDesign and evolve a modern LLVM/MLIR-based compiler stack.Develop compiler passes, lowering pipelines and backend code generation targeting custom GPU hardware.Optimise code generation, scheduling, memory access and overall execution efficiency.Improve end-to-end performance across compiler, runtime and framework integration.Work alongside hardware architects to influence future GPU capabilities through compiler design.Support CUDA-compatible execution for modern AI workloads across evolving hardware platforms.Mentor engineers and contribute to technical direction across the compiler organisation.What we're looking forStrong experience building production compiler infrastructure.Expert knowledge of LLVM and MLIR, including dialects, passes and backend development.Solid understanding of compiler optimisation, code generation and intermediate representations.Experience with GPU programming models such as CUDA and modern AI frameworks including PyTorch, JAX or ONNX.Background developing compilers for GPUs, AI accelerators or heterogeneous compute platforms.Strong C++ software engineering skills.Experience working closely with hardware teams or contributing to hardware/software co-design is highly desirable.