Global Capacity Manager

Operations
· San Francisco
· Remote
· Full Time
· Senior level

This role runs the supply and health of a large GPU fleet. That covers buying capacity, keeping clusters healthy, moving customer workloads between clouds, and building the software that makes it scale. A strong candidate has years of hands-on Kubernetes work, writes production Go or Python, and can model the cost of reliability against spend. Calm handling of outages and steady teamwork with SRE and infrastructure colleagues matter as much as the technical depth.

AI demand drives everything here, since customers' models need scarce, fast-changing chips, and the job is making sure those chips are available and used well. Automation can take over node triage, repair and workload moves, so the value shifts toward designing those systems and deciding what to build next. Judgment on supplier deals, cost trade-offs and incident calls stays with people.

The work puts a person close to how real AI inference runs at scale, including new hardware like Blackwell GPUs and how different models use capacity. Someone could build that background by running GPU workloads on Kubernetes, trying open-source model serving tools, and learning how GPU cost per request is worked out. Working with a platform that serves many AI companies gives a wide view of how the industry uses compute.

• Kubernetes operators in Go
• GPU or cloud capacity planning
• Cost and ROI modeling
• Incident response leadership

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

High-performance cloud inference platform for deploying, serving, and scaling open-source and custom artificial intelligence models.