Staff Software Engineer, Infrastructure Foundations

About the role

SUMMARY

Infrastructure Foundations builds and operates the multi-cloud platform underneath Temporal Cloud, turning the complexity of AWS, Google Cloud (GCP), and Azure into one cohesive platform the rest of Temporal builds on. Every cell we run is provisioned, managed, and safely changed by systems this team owns. When a customer's workflow runs, it runs on our work. We are heading into the next stage: Kubernetes-based infrastructure platforms built for the next orders of magnitude of growth.

We are looking for a Staff Software Engineer to design and build the next version of that platform. You will own a significant multi-component system, from the APIs other teams build on to how infrastructure state is modeled, changed, and recovered, and coordinate the engineers building its parts. You will stay close enough to the code to prove decisions out, build the capability roadmap for this area with the engineering manager, and lead through design reviews and the hardest incidents.

What makes this team different

- Real multi-cloud, not multi-cloud on a slide. We run across AWS, GCP, and Azure and keep extending the platform to new providers. The abstractions you design are the ones every new cloud plugs into.

- We build Temporal Cloud with Temporal. Cell creation, component deployment, and cloud provisioning run as Temporal workflows. You will shape how Temporal runs its own infrastructure on Temporal.

- A platform transformation, not maintenance. HA/DR for our control infrastructure, a declarative model for the whole platform, and a deployment control plane, all designed while a live platform keeps running.

- Everything sits on top of it. Nearly every team at Temporal depends on what we build, so the standards you set for this platform shape how those teams ship.

Responsibilities

- Multi-cloud architecture: Design the abstraction model that makes AWS, GCP, and Azure read as one platform. Decide where provider differences in IAM, networking, quotas, and failure modes get absorbed, and how new regions and providers plug in.

- Own a significant system: Design and build the multi-component platform end to end, and coordinate the engineers working on its parts. Make the system-level tradeoffs between conflicting requirements, and make them explicit.

- Declarative infrastructure model: Define how components are described, versioned, and reconciled across Kubernetes workloads and supporting cloud resources. Set the approach for reproducible environments, drift visibility, and safe reconciliation.

- Deployment control plane: Design the shared inventory, dependency awareness, rollout sequencing, and approvals that let many teams change production safely. Keep orchestration and safety policy cohesive while integrating with the frameworks that execute deployments.

- Reliability by design: Spot architectural flaws before they cost us performance or reliability. Make bounded blast radius, progressive rollouts, observability, load and failure testing, and recovery part of every design. Stay cool in the hardest incidents, then bring what we learn back into the platform.