Performance Modeling Engineer, AI Inference Deployments (Senior/Principal)

Role overview Build models and tools that estimate how large AI inference deployments will perform, including the hardware resources required and expected throughput and latency. The work combines analytical modeling, simulation, and measurements on real hardware. You will partner with customers, product, and chip modeling teams to turn deployment questions into clear, quantified guidance and a performance-estimation tool customers can run on their own workloads.

Responsibilities - Translate customer and product questions into focused studies with documented assumptions, options, metrics, and recommendations. - Compare equivalent deployments across accelerator platforms and explain the factors behind performance differences. - Model large-scale inference layouts using tensor, pipeline, and expert parallelism. - Build deployment-level performance models on top of chip simulators and validate estimates against hardware measurements. - Develop a standalone tool for customers to estimate performance on their own workloads. - Contribute to modeling architecture or per-model deployment analysis and customer-facing tooling, depending on the role focus.

Requirements - Experience in performance modeling, systems engineering, or performance analysis for machine-learning workloads. - Understanding of large-scale inference deployments and parallelism strategies. - Ability to combine analytical estimates, simulation results, and hardware measurements. - Strong communication skills for explaining assumptions, results, and trade-offs to customers and product teams. - Experience building tools that make technical analysis repeatable and usable by others.

Benefits and work setup The role is remote-first, with a company-provided computer and home office setup. Benefits include paid medical, dental, and vision coverage for employees and dependents, paid holidays, flexible paid time off, equity, and a retirement plan with company matching.