andromeda - Technical Program Manager - Provider Management

• Several years running technical programs in infrastructure. TPM or technical project management with genuine execution ownership, ideally involving external vendors, partners, or suppliers you didn't control. • Incident management experience: you've commanded or run point on production incidents involving multiple organizations, and you're comfortable with the off-hours reality that comes with that. • Enough technical depth to hold your own with a provider's data-center engineers and our SREs on GPUs, networking, and storage. You don't need to debug an InfiniBand fabric yourself, but you need to follow the conversation and know when someone is hand-waving. • A track record of getting teams you don't manage, including external partners, aligned and moving, including through rough patches where the relationship is strained. • Solid fundamentals: planning, risk tracking, status communication, and keeping four or five programs on the rails at once. • Calm, direct communication. You'd rather deliver bad news early than good news late, and both providers and internal teams trust you because of it. • Comfort with ambiguity. The process you'll follow mostly doesn't exist yet; you'll write a lot of it. • STRONG CANDIDATES MAY HAVE • Direct experience working with or inside neocloud, colocation, or data-center providers. • Vendor or supplier management background — SLAs, commitments, escalation frameworks. • Familiarity with the stack: NVIDIA data-center GPUs, InfiniBand/RoCE, Slurm or Kubernetes. • Experience supporting AI research labs or other large-scale GPU customers on the consuming side. • WHAT SUCCESS LOOKS LIKE • Within your first year, you'll have delivered: • Provider relationships we can bank on. Our compute providers trust you and deal straight with you, so when they tell us what capacity is coming and when, we believe it and plan against it. Capacity forecasting becomes reliable rather than hopeful. • Onboarding that's fast, predictable, and low-friction. New providers and sites come online on a known timeline with few surprises. Each onboarding is smoother than the last because you've systematized it. • A partner playbook that scales the network. You've built the standards and playbook that potential providers use to bring their infrastructure up to our bar. This makes qualifying and adding a new provider a repeatable process, not a one-off project every time. • Internal trust across the teams you don't manage. SRE, Engineering, Product, and Sales/CS rely on your read of provider health and timelines, and count on you to flag risk early and coordinate the response when it matters.