Group Product Manager, Compute Platform

Coupang•Mountain View, CA

3h•$190,000 - $265,000

About The Position

Coupang is at the forefront of the AI and high-performance computing (HPC) revolution. We are building a next-generation cloud platform designed to provide developers, researchers, and enterprises with seamless, scalable, and powerful access to accelerated computing. As the demand for AI, machine learning, and data-intensive workloads skyrockets, we are looking for a visionary product leader to define the future of our core compute offerings. CIC is looking for a Group Product Manager to own the foundational compute platform that powers enterprise AI workloads. This role spans fleet management, capacity management, bare metal, virtualized compute, node lifecycle, placement, reservations, and infrastructure-level customer experience. The Product Manager will be responsible for turning physical GPU and CPU capacity into reliable, customer-ready, observable, and billable compute products. This includes defining how capacity is reserved, provisioned, validated, monitored, maintained, packaged, and exposed to enterprise customers. This role sits at the compute foundation layer. It enables higher-level orchestration and workload services such as Kubernetes, Slurm, Ray, jobs, notebooks, and inference endpoints. The candidate should understand how those orchestration and workload systems depend on foundational compute infrastructure.

Requirements

8+ years of product management, technical product management, or equivalent product leadership experience in cloud infrastructure, compute, virtualization, GPU cloud, HPC, private cloud, or enterprise infrastructure platforms.
Experience working backwards from customer workload requirements to define infrastructure products, ideally for AI training, fine-tuning, inference, HPC, data-intensive workloads, or enterprise production systems.
Experience with foundational compute products such as virtual machines, bare metal, cloud instances, node pools, fleet management, capacity management, or infrastructure control planes.
Strong understanding of infrastructure concepts including provisioning, lifecycle management, placement, quota, reservations, OS images, networking, storage attachment, observability, and billing integration.
Familiarity with GPU-based infrastructure and the operational considerations that make compute capacity customer-ready, including drivers, firmware, OS images, high-performance networking, and workload compatibility.
Ability to translate customer AI workload needs, such as distributed training, inference serving, data movement, checkpointing, and cluster operations, into product requirements for compute capacity, lifecycle, observability, and enterprise readiness.
Experience partnering with engineering and infrastructure teams on technically complex systems while driving product outcomes, roadmap decisions, prioritization, and business impact.
Experience supporting enterprise customers, including workload discovery, requirements definition, launch readiness, customer-facing documentation, GTM enablement, and post-launch adoption measurement.
Strong analytical judgment around workload requirements, utilization, capacity planning, product readiness, revenue impact, support cost, and customer adoption.
Strong written and verbal communication skills with engineering, infrastructure operations, finance, sales, support, executive stakeholders, and enterprise customers.

Nice To Haves

Experience at a hyperscaler, neo-cloud, GPU cloud provider, HPC cloud provider, private cloud platform, or infrastructure SaaS company.
Experience with NVIDIA GPU infrastructure, CUDA, NCCL, OFED, driver compatibility, firmware lifecycle, GPU health/telemetry, or supported GPU software stack management.
Experience with distributed AI training or inference infrastructure, including Slurm, Kubernetes, Ray, model serving platforms, high-performance networking, shared storage, or checkpointing workflows.
Experience with reserved capacity, committed-use contracts, dedicated clusters, savings plans, or enterprise infrastructure commitments.
Experience with bare metal-as-a-service, GPU passthrough virtualization, VM image lifecycle, custom images, or cloud control plane products.

Responsibilities

Define the product strategy and roadmap for CIC's compute platform across fleet, capacity, bare metal, and virtualized compute, working backwards from customer AI workload requirements.
Own the product lifecycle from infrastructure capacity to customer-ready compute, including reservation, provisioning, lifecycle actions, observability, billing integration, maintenance, and deprecation, with clear linkage to customer workload readiness.
Define customer-facing abstractions for capacity, node pools, bare metal instances, VM instances, placement policies, OS/runtime images, and reserved compute based on how customers run AI training, inference, and cluster operations.
Partner with engineering and infrastructure operations to define fleet readiness, lifecycle states, failure handling, maintenance workflows, and operational requirements.
Partner with sales, solutions, support, and finance to understand enterprise AI workload requirements and translate them into compute offerings, including reserved capacity, dedicated infrastructure, and VM/bare metal packaging.
Define product requirements for supported compute configurations, including GPU type, node shape, OS/runtime image, network/storage attachment, and compatibility expectations for customer AI workloads.
Improve the reliability, usability, and supportability of CIC compute products across customer onboarding, provisioning, and day-2 operations.
Drive roadmap decisions that improve utilization, reduce stranded capacity, increase enterprise readiness, and lower operational support burden.
Support enterprise customer discovery and roadmap prioritization for bare metal, VMs, reserved capacity, and dedicated infrastructure, with specific attention to customer workload patterns, performance needs, operational workflows, and production readiness.
Create clear product narratives, customer-facing materials, sales enablement, launch plans, and executive updates for enterprise compute offerings.
Track product and business metrics such as usable capacity, allocated capacity, stranded capacity, provisioning time, replacement time, utilization, revenue per deployed GPU, and support burden.