Compiler Engineer

Compiler Engineer - Up to $350k + Equity - Hybrid I’m supporting an AI infrastructure company developing compiler technology that allows machine-learning models to achieve strong performance across multiple compute architectures. This role covers the journey from model-level representations through optimization and into machine-code generation. You’ll help determine how AI workloads are transformed, scheduled and mapped efficiently onto GPUs and other accelerators. Core responsibilities include: Creating graph and intermediate-representation optimization passesImplementing fusion, tiling, layout transformation and algebraic simplificationDeveloping target-aware instruction selection and register allocationImproving loops through vectorization, parallelization and unrollingBuilding automated kernel-generation capabilitiesWorking with linking, relocations, calling conventions and object formatsProfiling compiled workloads to identify latency and throughput improvementsPartnering with runtime and hardware engineers on architectural decisions The ideal background: At least four years developing compiler technologyAdvanced C or C++ skillsStrong foundations in compiler algorithms and data structuresExperience with IR design, transformation passes or backend code generationKnowledge of GPU, accelerator or heterogeneous-compute architectures Exposure to LLVM, MLIR, XLA, PyTorch internals, affine analysis, quantization or machine-learning workloads would be advantageous. You’ll have genuine ownership over technology that directly controls how efficiently modern AI models execute.