Global Capacity Manager – TPU Focus

Jobtailor; California (MO); On-site; USD 180,000 - 240,000; Full time

Job summary

Jobtailor is seeking a senior TPU capacity engineer to lead TPU pod fleets, design capacity management, and optimize multi-cloud environments for cost and performance.

The role requires deep Kubernetes and GCP TPU expertise, Go or Python production experience, and strong collaboration across teams to secure enterprise TPU capacity.

Qualifications

• Bachelor's, Master's, or Ph.D. degree in Computer Science, Engineering, Mathematics, or a related field.
• 5+ years of professional work experience in a high-growth environment, preferably at a hyperscaler (GCP, AWS, Azure) or a specialized accelerator provider.
• Hands-on experience with Google Cloud TPUs, including pod slicing, ICI topology, JAX/XLA, and TPU-specific scheduling and fault handling
• Deep expertise in Kubernetes, including taints, cordons, node draining, and custom operators
• Demonstrated experience with Go or Python in a production-level environment
• Strong financial literacy and ability to model complex trade-offs between capacity reliability and cost
• High tenacity and collaborative mindset
• Experience with additional non-NVIDIA accelerators, such as AWS Trainium/Inferentia or AMD Instinct, is nice to have
• Familiarity with multi-accelerator scheduling and cost/performance tradeoff modeling is nice to have
• Prior experience partnering with model performance or ML systems teams is nice to have

Responsibilities

• Lead TPU pod fleets through acquisition, allocation, and maintenance
• Execute complex workload migrations and sticky deployment drains across TPU topologies
• Ensure deployment scheduling rules meet regional and compliance requirements
• Design and implement the next version of Baseten's capacity management system
• Scale TPU capacity alongside the existing GPU fleet
• Build ROI models comparing TPU, GPU, and other accelerator options
• Partner with MP, SRE, Infra, and FDE teams to tune workloads for TPU execution
• Lead capacity-crunch incident response by reallocating and re-coordinating TPU workloads
• Architect infrastructure readiness and deployment strategies for TPU clusters, including pod slicing and topology planning
• Build multi-cloud capacity management systems to optimize cost and latency
• Develop automated operators to identify, cordon, and repair unhealthy TPU pods
• Partner with leadership and Google Cloud to secure dedicated TPU capacity for enterprise customers

Skills

Google Cloud TPU Management

Kubernetes Expertise

Go or Python Programming

Financial Modeling

Collaborative Mindset

Experience with Non-NVIDIA Accelerors

Multi-Accelerator Scheduling

Model Performance & ML Systems Liaison

TPU Pod Slicing

ICI Topology

JAX/XLA

Custom Operators

Node Draining

Education

Bachelor's/Master's/Ph.D. in Computer Science, Engineering, Mathematics, or related field

Tools

Google Cloud

AWS

Azure

Multi-Cloud Systems

Job description

• Lead TPU pod fleets through acquisition, allocation, and maintenance
• Execute complex workload migrations and sticky deployment drains across TPU topologies
• Ensure deployment scheduling rules meet regional and compliance requirements
• Design and implement the next version of Baseten's capacity management system
• Scale TPU capacity alongside the existing GPU fleet
• Build ROI models comparing TPU, GPU, and other accelerator options
• Partner with MP, SRE, Infra, and FDE teams to tune workloads for TPU execution
• Lead capacity-crunch incident response by reallocating and re-coordinating TPU workloads
• Architect infrastructure readiness and deployment strategies for TPU clusters, including pod slicing and topology planning
• Build multi-cloud capacity management systems to optimize cost and latency
• Develop automated operators to identify, cordon, and repair unhealthy TPU pods
• Partner with leadership and Google Cloud to secure dedicated TPU capacity for enterprise customers

Requirements

• Bachelor's, Master's, or Ph.D. degree in Computer Science, Engineering, Mathematics, or a related field
• 5+ years of professional work experience in a high-growth environment, preferably at a hyperscaler (GCP, AWS, Azure) or a specialized accelerator provider
• Hands-on experience with Google Cloud TPUs, including pod slicing, ICI topology, JAX/XLA, and TPU-specific scheduling and fault handling
• Deep expertise in Kubernetes, including taints, cordons, node draining, and custom operators
• Demonstrated experience with Go or Python in a production-level environment
• Strong financial literacy and ability to model complex trade-offs between capacity reliability and cost
• High tenacity and collaborative mindset
• Experience with additional non-NVIDIA accelerators, such as AWS Trainium/Inferentia or AMD Instinct, is nice to have
• Familiarity with multi-accelerator scheduling and cost/performance tradeoff modeling is nice to have
• Prior experience partnering with model performance or ML systems teams is nice to have

Core Competencies

Demonstrates expertise in managing TPU pod fleets, including acquisition, allocation, and maintenance, while ensuring compliance with deployment scheduling rules. Proficient in designing capacity management systems and optimizing multi-cloud environments for cost and performance.

Highest-signal resume keywords

• Google Cloud TPU Management
• Kubernetes Expertise
• Go or Python Programming
• Capacity Management Systems
• Financial Modeling

Hard Skills

• TPU Pod Slicing
• ICI Topology
• JAX/XLA
• Custom Operators
• Node Draining

Soft Skills

• Collaborative Mindset
• High Tenacity

Industry Keywords

• Hyperscaler
• Accelerator Provider
• Capacity Reliability
• Cost/Performance Tradeoff

Tools & Technologies

• Google Cloud
• AWS
• Azure
• Multi-Cloud Systems