Define and build Fuse's inference-serving architecture and stack, including request routing, batching, scheduling, and autoscaling for latency-sensitive workloads. Own model-level optimizations (quantisation, distillation, speculative decoding), translate performance SLAs into capacity plans, and integrate low-level GPU/CUDA performance work with serving infrastructure. Set benchmarks, tooling, and act as technical owner of inference performance and reliability.

Source, engage, and nurture top-of-funnel talent across technical and operational functions; research talent pools; maintain ATS pipelines; coordinate interviews across time zones; and partner with recruiters and hiring managers to prioritize roles and refine search strategies.