Project Manager (Enterprise GPU Cluster Projects)
Own enterprise GPU cluster projects end to end, from deal readiness to go-live, as the first Project Manager at a GPU cloud startup growing several-fold a year.
About The Job
This role is open in Seoul right now. Harper is an AI career agent that works for you — the candidate. Skip the application form: just talk to Harper about where you are and what you want next, and Harper connects you with this role and other frontier roles worth your time.
Job Title
Project Manager (Enterprise GPU Cluster Projects)
Location
Seoul (Gangnam), on-site with flexible start times (8–11am). Full-time.
Company Description
A Seoul-and-San-Francisco GPU cloud ("Neo Cloud") startup that gives AI teams the GPU compute they need, when they need it, by connecting GPU capacity from multiple cloud and data center partners into one cloud. Founded in 2020, ~₩22B raised, 100+ customers including a top automaker's autonomous-driving team, a leading mobility platform's AI-agent team, academic and frontier labs, and large public-sector AI programs. Revenue grew 3.4x year over year.
The Role
This is not a project coordinator who tracks schedules. You are the single-threaded owner of enterprise GPU cluster projects: you turn customer requirements into an executable plan and own everything from deal readiness through contract, build-out, acceptance, go-live and handoff to operations. You compare customer RFIs against what the company can actually deliver, check that the GPU count, timeline, networking and SLA in the contract match real supply agreements, and stop a commitment or go-live when a critical gap is still open. You are the first Project Manager on the Korea GTM team, working with AEs, Forward Deployed Engineers and technical owners across infra, network and platform.
What You'll Do
-
Deal readiness: with the AE before close, structure the customer's requirements (RFI) and run fit/gap analysis against the delivery scope, supply-contract volumes, timelines and SLAs
-
Technical consultation: discuss workloads, cluster configuration (compute, network, storage) and operating model with the customer's engineers and propose a configuration within standard scope
-
End-to-end delivery: from kick-off to build-out, acceptance, go-live and handoff, own the plan, milestones and dependencies
-
Risk & escalation: catch schedule, supply, technical and security risks early; separate standard scope from non-standard asks and bring their cost, timeline and resource impact to leadership
-
Stakeholder governance: run regular reviews with customer teams and executives, cloud and data center partners, telcos, MSPs, and internal FDE, Product, Engineering and Infra teams
-
Go-live & handoff: confirm contract, infra, network, security and acceptance readiness; onboard the customer and hand operations to Support, Infra or a designated MSP
-
Delivery process: build the RFI template, deal readiness review, go-live checklist, acceptance criteria and retrospectives, and turn repeated asks into standard services and playbooks
Who You Are
-
5+ years running complex customer projects in cloud infrastructure, SI/MSP, network or storage vendors, or technical professional services
-
You have led customer-facing projects A to Z: requirements, build-out, acceptance and handoff
-
You understand GPU cluster building blocks (GPU servers, high-performance networking, storage, Kubernetes or Slurm) well enough to discuss trade-offs with customers and engineers
-
You have managed scope, schedule and risk across customers, supply partners and several internal teams, and tied SOW/SLA scope to the actual delivery plan
-
You can structure incomplete information and conflicting requirements, and explain status and risk clearly to customer executives, engineers and leadership
Even better
-
GPU/HPC cluster or AI infrastructure build-out or operations projects
-
Large enterprise projects at a cloud provider, neo cloud, MSP or SI
-
Multi-party projects with data centers, telcos, cloud providers, hardware vendors and MSPs
-
Familiarity with procurement, security review, contracting and acceptance at large Korean enterprises; finance, public sector or bio projects with strict compliance
-
Business-level English
Process
Resume interview (about 1 hour) → deep-dive and culture interview (up to 3 hours) → CEO interview → reference check. Full-time with a 3-month probation period.
Why Join
-
Lead GPU cluster projects of hundreds to thousands of GPUs end to end, and see the whole AI infrastructure business, from supply contracts to customer workloads
-
Design the delivery system as the first Project Manager on the Korea GTM team, working directly with customer executives and the company's C-level
-
Annual US market exposure trip, self-development budget, health checkups, flexible hours, long-service refresh leave