Staff+ Software Engineer, Platform Portability

Anthropic•San Francisco, NY

1d•Hybrid

About The Position

Anthropic trains and serves Claude across multiple cloud providers and accelerator platforms. The Platform Portability team owns the abstractions, tooling, and deployment infrastructure that make this possible without every engineering and research team needing to become experts in different clouds. As an engineer on Platform Portability, you'll design and build the systems that let us bring up new capacity, ship our platform to cloud provider platforms & marketplaces, and move workloads between providers with confidence. You'll work closely with our cloud partners on capabilities we need that don't exist yet, and with internal Compute, Inference, and Security teams to make "which cloud is this running on?" a question most engineers at Anthropic never have to ask. Your work directly determines how quickly and reliably we can put safe, frontier AI in front of users — wherever they are.

Requirements

A minimum of 8 years of practical experience building and operating production distributed systems.
Experience owning large-scale infrastructure systems and leading ambiguous, cross-team technical projects from design through production.
Experience running the same production workload across multiple cloud providers, including migration or abstraction work.
Understanding of how cloud IAM models and networking primitives (VPC peering, private connectivity, service mesh) differ across providers.
Proficiency in writing production code in at least one general-purpose programming language (e.g., Python, Go, or Rust).
Work confidently with Kubernetes and Infrastructure as Code (e.g., Terraform, Pulumi).
Understand ML infrastructure and supporting interconnects.
Experience shipping software through hyperscaler marketplaces or managed-model platforms.
A background in security, compliance, or data residency for regulated or sovereign cloud environments.
Care about the societal impacts of your work and Anthropic's mission to build safe, beneficial AI.

Responsibilities

Design and build provider-agnostic abstractions for storage, messaging, and connectivity that allow Anthropic's systems to run on any cloud at the scale of frontier model inference.
Develop tooling that enables internal teams to provision, configure, and operate services across clouds through a single consistent interface.
Identify and eliminate provider-specific assumptions in the codebase and establish patterns to prevent new ones.
Own the end-to-end delivery of shipping Claude onto third-party cloud platforms, including build, release, and validation pipelines.
Collaborate directly with engineering teams at cloud partners and internal infrastructure stakeholders to implement necessary features, debug platform issues, and influence roadmaps.
Gain deep expertise in native integrations with each cloud provider and perform cross-cutting work to maintain platform consistency across all providers.
Shape how the rest of Anthropic builds for multiple clouds by reviewing designs with cloud-portability implications and mentoring engineers on multi-cloud patterns.
Align internal stakeholders through clear writing and conversation.
Participate in the on-call rotation and design postmortems, runbooks, and incident response procedures.
Hunt down the provider-specific assumptions hiding in our codebase and set the patterns that keep new ones from creeping in.