Lead Inference Engineer

Plexus has partnered with a rapidly growing AI company looking for a GPU Performance Engineering Lead to build and lead a specialist infrastructure team. This is a hands-on leadership role focused on improving the speed, efficiency and reliability of large-scale AI workloads. You’ll set the technical direction, hire the team and work directly on complex performance challenges across distributed GPU infrastructure.

Responsibilities: Own the technical strategy for LLM inference performance Recruit and lead the Inference Optimisation team Optimise latency, throughput and cost per token across GPU infrastructure Benchmark inference engines, quantisation methods and parallelism strategies Improve inference routing and load balancing Evaluate new kernels, hardware and optimisation techniques Requirements: 8+ years in performance optimisation, HPC or GPU engineering 5+ years leading engineering teams Production experience with vLLM, SGLang or similar inference engines at scale Strong knowledge of GPU architecture, distributed inference and quantisation Experience with GPU profiling tools such as Nsight or PyTorch Profiler Proficiency in Python, Rust or Go C++, CUDA or Triton experience is highly desirable Offer: $270k–$330k base salary Fully remote within the US Opportunity to build and lead the inference function Work on high-volume AI infrastructure and cutting-edge GPU technology If this piques your interest, please apply now!