Performance Engineer

Overview

Cuspai is seeking a Performance Engineer to connect frontier AI models with high-performance computing infrastructure. The position involves building optimization layers and profiling infrastructure to maximize the efficiency of large-scale training and inference workloads. The role operates on a hybrid schedule with time spent in the office three days per week, and may involve regular travel to other locations for project collaboration.

Who can apply

Candidates must have at least three years of experience in performance engineering, site reliability engineering, or low-level software development focused on high-performance computing or AI workloads. Applicants are required to have expert-level knowledge of GPU architecture, experience optimizing CUDA kernels or using libraries like Triton, and a deep understanding of distributed training frameworks and networking protocols. Proficiency in profiling tools such as Nsight Systems, Nsight Compute, or Weights & Biases is also mandatory.

What you get

The position offers a competitive salary along with company equity. Employees receive 28 days of holiday in the UK, Germany, and the Netherlands, plus local public holidays. Parental leave includes 26 weeks at full pay for primary caregivers and 12 weeks at full pay for secondary caregivers. Additionally, a professional development budget is provided to support ongoing career growth.