Technical Product Manager

The AI Compute Platform runs the GPU infrastructure frontier AI labs train and serve their models on. The Compute API is how customers — and a dozen internal product teams — actually use it. As the TPM for Compute API & Experience, you own that API end-to-end and the surface customers touch — console, CLI / SDK, instance metadata, self-service and the reliability signals they trust us on. It's a gatekeeper seat with real authority: you set the contract other teams build on, and own the VM lifecycle and scheduling logic that decide how customer VMs are placed, run and recovered — on one of the largest GPU fleets in Europe. Responsibilities: Own end-to-end product responsibility for your area — strategy, roadmap, discovery, delivery, adoption, and measurable customer & platform outcomes. Design and govern platform contracts at hyperscaler quality. Co-design the compute control plane with engineering as a technical peer (scheduling, allocation, reconciliation) — not just word an API. Manage stakeholders and drive cross-team execution across engineering, networking, storage, product and sales / CX. Define success metrics for the Compute API and the customer experience, and be the escalation point for product decisions on your surface. We expect you to have: 6+ years in Product / Platform / Infrastructure PM — or an SRE / Engineering Lead moving to product — shipping technically complex platform products with measurable impact. Owned a public cloud / compute / platform API as a product. A declarative / desired-state or control-plane API (Kubernetes CRDs / operators, or a cloud control plane) is a strong plus. Hands-on cloud experience — the VM / instance lifecycle, its API and its console at a public or large-scale cloud (AWS / GCP / Azure or another large-scale cloud), with a real under-the-hood understanding of how scheduling, allocation and the virtualization layer work (control-plane vs data-plane — how VMs are placed, scheduled and run). GPU / AI-cloud experience is a big plus, but not required. A systems understanding of how a cloud fits together — how compute, storage, networking, IAM and related services connect and interact, and how it all works under the hood — enough to design the Compute API coherently and reason about it with those teams. Gatekeeper craft — other teams have shipped features through an API you governed: design review, breaking-change control, pushback with a migration path; debates trade-offs with engineering as a peer. Strong analytical skills — comfort defining and instrumenting product metrics, working with telemetry, and building data-informed roadmaps. Experience leading discovery-heavy work — structured customer interviews (we build top-tier infrastructure and work directly with frontier AI labs), usage analytics — turning insights into shipped product. Strong communication and the ability to align engineering, SRE, customer-facing teams and exec stakeholders. High ownership, a bias to ship, and focus on outcomes and customer value. Nice to have: API & developer-surface depth, System design of compute / platform services — control plane, scheduling, allocation, reconciliation, Observability / SLI-SLO delivered as a product; status pages / RCA / SLA artefacts. Console / CLI / API design and developer-experience product craft. Hardware, GPU & HPC depth, Hands-on with GPUs / accelerated compute and ML / AI workloads (hands-on ML experience). Multi-node GPU clusters and their interconnect — InfiniBand / RoCE fabrics, NVLink topology, topology-aware placement. NVIDIA reference architectures; brought a new instance type / preset / hardware platform into a cloud's compute catalog. Systems depth, Understanding of virtualization internals (hypervisors / KVM / QEMU, VM internals). Hands-on experience with Kubernetes.