Staff+ Software Engineer, Deployment

About the role

Every change to Anthropic's software has to reach production somehow: new code, runtime configuration, feature flags, infrastructure changes. Our team builds the systems those changes travel through. Our mission is to give every team at Anthropic a fast, safe, and predictable path from change to production.

We sit within Anthropic's Infrastructure organization and own pipelines, staged rollouts, health-gated promotion, and rollback, along with runtime configuration, feature flags, and the systems that deliver code and artifacts to machines across a very large fleet. You'll own a meaningful part of this end to end, from design through operating it in production.

If you want your work to shape how fast and how safely Anthropic ships, this could be a great fit.

Key responsibilities

• Set technical direction for major parts of the deployment platform and own them end to end: design, implementation, and operations
• Build rollout safety into the platform, including progressive delivery, health-gated promotion, automatic pause and rollback, and deploy controls that teams manage themselves
• Design APIs and abstractions that work equally well for engineers and for AI agents
• Extend the platform to new workload types and runtimes, from Kubernetes to serverless and edge platforms
• Build and operate runtime configuration and feature flag systems with validation, gradual rollout, and fast rollback
• Build infrastructure-as-code workflows that many teams rely on for safe, routine changes
• Scale artifact distribution so code and configuration reach machines quickly and efficiently
• Partner with Security so that the path to production is secure by default, with verifiable builds, least-privilege access, and auditable changes
• Work directly with teams across Anthropic to understand their deployment needs and bring their workloads onto the platform
• Take part in on-call and incident response, and turn what you learn into guardrails that prevent the same problem from recurring

Minimum qualifications

• Proficiency in at least one backend or systems programming language, such as Rust, Go, C++, or Java
• Experience leading the design, building, and operation of distributed systems or infrastructure in production, including technical direction across teams
• Experience with cloud infrastructure and containerized workloads, such as Kubernetes on AWS or GCP
• Strong written and verbal communication skills

Preferred qualifications

• Experience building deployment, release, or change management systems at large scale
• Experience with production on-call and incident response
• Experience with Rust in production
• Experience with progressive delivery and GitOps tooling, such as Argo CD or Argo Rollouts
• Deep knowledge of Kubernetes, including controllers and operators
• Experience with Terraform or other infrastructure-as-code systems at scale, including state management and locking
• Experience building feature flag, experimentation, or dynamic configuration systems
• Experience with artifact or package distribution, such as OCI registries, Nix, or content-addressed storage
• Experience with software supply chain security, such as signed build provenance
• Experience building high-throughput workflow engines or control planes backed by relational databases
• Experience building internal developer platforms and leading adoption across many teams
• Experience using AI agents in your own engineering work, or building systems that agents operate

Representative projects

• Build operator controls for progressive rollouts, including pause and resume that preserve rollout state
• Design a deployment engine built on fine-grained targets and their dependencies, so each change rolls out exactly where it needs to
• Build cluster-health-aware rollouts that steer around unhealthy clusters
• Add a new workload type to the platform, including its rollout, verification, and rollback behavior
• Design a deployment control plane that handles sharp bursts of agent-driven traffic
• Design a distribution system that delivers configuration and artifacts to every machine in the fleet as demand grows