Compiler Optimisation Engineer

Compiler Optimisation EngineerWe are working with an ambitious AI infrastructure company building a new software stack that enables machine-learning workloads to run efficiently across diverse hardware targets.They are looking for a Compiler Optimisation Engineer to own the graph-optimisation layer of their AI compiler: the point where high-level model graphs are transformed into efficient execution plans before code generation.You will work on operator fusion, layout propagation, simplification and IR design, with a direct, measurable impact on model latency and throughput across GPUs and other accelerators.What you’ll doDesign and implement graph-level optimisation passes for a heterogeneous AI compilerWork on fusion, layout optimisation, dead-code elimination, constant folding and algebraic simplificationEvolve the compiler IR as model architectures and hardware targets developUse performance analysis to identify optimisation opportunities and improve execution efficiencyPartner closely with frontend and code-generation engineers on clean compiler interfaces and robust validationWhat we’re looking for4+ years of compiler engineering experience, particularly IR design or optimisation passesStrong C/C++Deep understanding of graph and tensor optimisation, including fusion, tiling and layout transformationsExposure to MLIR, XLA or comparable compiler frameworks would be valuableFamiliarity with PyTorch, JAX, TensorRT, GPUs or AI accelerators is beneficialAn interest in solving difficult performance problems at the intersection of ML models, compilers and hardwareThis is a chance to own a critical part of a next-generation AI compiler, working on infrastructure that makes advanced models portable and performant across hardware. You will join a technically serious team early enough to shape both the architecture and direction of the platform.