Overview:
In this role you own the technical success of GPU/CPU cluster deployments, leading implementation, integration, and optimization from provisioning to steady-state operations. You will serve as the primary technical contact across customer accounts and work with cross-functional teams to ensure high-performance, reliable compute environments. The position offers ownership in a fast-paced, high-expectation setting that powers AI capabilities for a strategic partner and SpaceXAI’s mission. You’ll influence end-to-end delivery, risk management, and continuous optimization, shaping scalable infrastructure for demanding workloads.
Responsibilities:
Lead technical implementation, integration testing, provisioning, and go-live for GPU/CPU clusters and supporting infra. Define compute infrastructure requirements and serve as the primary technical contact across customer accounts. Perform design validation, performance tuning, and optimization of large-scale compute systems. Provide deep expertise on GPUs/CPUs, high-performance networking, data center connectivity, and related systems. Handle escalations, root-cause analysis, and ensure SOW/SLA compliance. Coordinate with Hardware, Network, Facilities, Cloud Compute, SRE, and PM teams for end-to-end delivery. Engage in verification testing, performance characterization, reliability assessments, and continuous optimization. Support capacity planning, expansion, and new use cases across customers. Ensure on-time deliverables and proactive risk mitigation to meet mission goals.
Key requirements:
Bachelor’s degree in a relevant technical discipline. 7+ years of hands-on experience in large-scale compute, data center, or hyperscale environments. Experience leading technical implementations involving GPUs/CPUs and high-performance systems. Excellent written and verbal communication skills. Ability to interface with customers and internal stakeholders effectively. Strong ownership mindset and capability to work in a fast-paced, high-expectation environment. InfiniBand RoCE NVIDIA networking technologies (NCCL, NVLink, GPU Direct RDMA)