SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 285,000 - 0 / annual
Crusoe is a vertically integrated AI infrastructure company that owns and operates the complete stack—from energy generation to AI workloads—to power the world's most ambitious AI compute. The company is solving the energy bottleneck constraining AI infrastructure growth through an energy-first approach.
As Principal Product Manager for AI Infrastructure (Networking), you will define Crusoe's 2-4 year networking vision and roadmap, owning the complete networking foundation from fleet-wide architecture to fabric design. You will establish the long-term vendor management strategy, defining cost structure and capital allocation for networking hardware and services. You'll set performance, reliability, scalability, and cost targets—including network cost per GPU and per server—that the infrastructure must achieve.
Key responsibilities include owning cross-datacenter and multi-site connectivity strategy, network topology design, and failure-domain architecture. You will drive the technology strategy for NICs, DPUs, switches, and networking hardware, evaluating new generations and vendors to inform long-term investments. You'll identify structural limitations in the current architecture and prioritize foundational investments to resolve them.
You will partner closely with Infrastructure Engineering and Architecture teams on major technical tradeoffs, defining product requirements and business rationale for investments. You'll represent networking foundations in leadership reviews, translating complex architecture and vendor decisions into clear decision documents for executives.
You bring 8+ years of technical product management or product-minded infrastructure engineering experience at Principal scope, with deep subject matter expertise in networking. You have direct experience with fleet-wide network architecture, topology design, and failure-domain strategy at scale. You understand infrastructure economics, including cost per unit and capacity planning. You have hands-on experience with RDMA fabrics (InfiniBand or RoCE) and GPU cluster networking for AI training or inference. You communicate exceptionally well, translating technical complexity into clear narratives for both engineers and executives.
