Principal Performance Modeling Architect

Principal Performance Modeling ArchitectJoin a well-funded semiconductor startup developing next-generation compute technology for large-scale AI workloads. The company is building across silicon, systems software and infrastructure to improve the performance and economics of AI inference and training.As Principal Performance Modeling Architect, you’ll own the platform used to evaluate proposed accelerator and cluster architectures before hardware is built. Your work will produce defensible projections for performance, energy efficiency and system cost, directly influencing technical and commercial decisions.This is a hands-on senior individual-contributor role with significant ownership of the platform’s technical direction.Key requirements:5–8+ years in performance modelling, computer architecture or systems performanceExperience building analytical, first-principles or roofline models for AI/HPC systemsStrong knowledge of GPUs or AI accelerators, HBM, memory hierarchies and interconnectsUnderstanding of distributed LLM inference/training and parallelism strategiesStrong Python and experience developing maintainable modelling softwareAbility to validate simulation results critically and translate findings into architecture decisions