> Markdown version of [/jobs/ext/4579993-senior-program-manager-deployment-automation](https://www.wearedevelopers.com/jobs/ext/4579993-senior-program-manager-deployment-automation). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Senior Program Manager, Deployment Automation - **Company:** Nscale AS - **Location:** New York, NY, United States - **Experience:** Expert - **Salary:** $160,000.0 - $300,000.0 - **Contract:** Temporary contract - **Skills:** Artificial Intelligence, Computer Clusters, Data Centers, Network Configuration and Change Management, Graphics Processing Unit (GPU), Computer Network Operations, Hardware Testing, Deployment Automation - **Published:** November 5, 2026 - **Apply:** https://www.careerbuilder.com/job-details/senior-program-manager-deployment-automation-new-york-ny--1d1a7f4f-fcc8-4c94-a32b-31774c4bc7b7 ## About the Role * 7+ years in data center, hyperscale infrastructure or large-scale hardware deployment, with substantial time on site. * Hands-on delivery of GPU or HPC clusters, from data hall readiness through rack energization. * Working knowledge of data center power, liquid and air cooling, structured cabling and rack integration. * A track record of holding OEMs, data center operators and construction partners to firm commitments. * Strength in building and running detailed site schedules and critical paths, and in recovering them when they slip. * Clear, concise communication, in writing and on site, with honest forecasts. * Ability to travel 75% or more of the time, primarily to data center sites, often at short notice., * Hands-on experience with NVIDIA GB200/GB300 or Vera Rubin NVL72-class rack deployments. * Experience at a hyperscaler, neocloud or AI lab deploying 10K+ GPU clusters. * Familiarity with rack-level burn-in and hardware validation. * Experience using tooling or automation for rack tracking and deployment reporting. * PMP, or a construction or engineering background. ## Description The Onsite Technical Program Manager is Nscale's owner on the data center floor. You drive a cluster from data center early access until racks are energized and ready for logical bring-up, and you track the state of every rack along the way. You turn construction, power, cooling and supply chain realities into a daily plan, and you hold partners to it. Your scope starts when the data hall is handed to Nscale and ends when every GPU rack is powered on, cabled and passing hardware acceptance testing. It covers site readiness, rack receipt and placement, power and cooling activation, structured cabling, and vendor-led hardware testing. You hand a clean, documented floor to the Logical Deployment team for network configuration, validation and customer handoff. This is an individual contributor role reporting to the Director, Physical Deployment. It is a field role, so expect to spend 90% or more of your time on site, working from the data center floor. What you'll do Own the site schedule * Own the physical schedule for your site: data center early access, rack delivery, power, cooling, cabling and hardware acceptance testing. * Track and defend milestone dates. Flag early when float disappears, with a recovery option ready. * Run daily stand-ups on site with data center partners, OEMs (such as Dell, Lenovo and Supermicro), cabling vendors and onsite crews. * Keep rack-level status current, so leadership can see exactly where every rack stands. Manage supply and site risk * Track GPU, rack, switch, transceiver and cabling material arriving on site against need dates, with Supply Chain and bill of materials (BOM) specialists. * Work with data center partners on power and cooling readiness, including primary cooling loops, phased energization and temporary solutions. * Drive recovery on site when supply, construction or vendor staffing slips, re-sequencing work, re-routing hardware and escalating for resources. Hold partners to the plan * Verify vendor work against statements of work (SOWs), staffing plans and quality standards, and raise gaps the same day. * Walk the floor daily and confirm progress first-hand, not from vendor reports alone. * Hand off cleanly to Logical Deployment, with complete documentation, punch lists and test results. Improve how we deploy * Follow the physical deployment playbook, and feed back what works and what doesn't after every site. * Help the tooling team shape automation for rack tracking, cabling verification and status reporting. * Report daily and weekly status and risks to the Director, Physical Deployment, clearly separating forecast from upside. What success looks like in your first 6-12 months * Your site runs on one integrated physical schedule, reviewed daily on the floor and weekly with leadership. * Physical milestones hit their committed dates, or slips are flagged at least two weeks ahead with a recovery plan. * Rack-level status for your site is current and visible without manual chasing. * Logical Deployment receives your floors with no open blockers at handoff. * You have contributed at least one improvement to the physical deployment playbook., This is a leader who owns outcomes, not one who reports on them. * Decisive: You call the real date, even when it's unpopular, and put a recovery option next to it. You make calls on the floor with incomplete information and stand behind them. * Assertive with partners: You hold data center partners, OEMs and vendors to their commitments. When a partner slips, you escalate early, directly and with evidence. * Accountable: You own the outcome at your site. Rather than "waiting on the vendor," you explain what you are doing about it. * Steady under pressure: You bring order to a site in crisis and keep crews focused when the schedule is tight. * Credible: You brief leadership and customers with a clear, honest position: what's true, what's at risk, and what you need. * Hands-on: You walk the data hall and know the state of the floor first-hand. Status reports alone are not enough.