Staff+ Software Engineer, Platform Connectivity
A Staff Software Engineer on the Platform Connectivity team at Anthropic, designing, building, and operating core connectivity infrastructure (MCP proxy, OAuth, token management) for AI agents, ensuring reliability and enterprise trust, contributing to open standards and SDKs.
- Compensation
- $405K–$485K / yr
- Location
- London, UK; San Francisco, CA
- Work setup
- Not specified
- Seniority
- Staff
- Employment
- Not specified
- Experience
- Not specified
- Category
- Software engineering
- Posted
Skills
Must have
- Experience as a backend or platform engineer
- Familiarity with using AI tools as part of the software development process
- Take a product-focused approach to platform work and care about building solutions that are robust, scalable, and easy to use
- Take full ownership of your work — from design through build, deployment, and operations, to iterating on and improving your product
- Can navigate ambiguity and make sound technical decisions independently, and have ideally operated in both 0 → 1 and more mature team or company settings
- Experience operating production distributed systems, with an understanding of what reliability, observability, and incident response require at scale
- Excellent communication skills, with the ability to build consensus across teams and time zones
Full description
About Anthropic
Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems.
About the role
Our mission on the Connectivity team is to make Claude the most connected AI. As agents take on more real-world work, we're taking a leading role in defining how they get access to external systems — safely, with the right permissions, and in ways enterprises can trust. We ship standards and services used by claude.ai, Claude Code, Cowork, and the API. We own the MCP proxy that routes every tool call and the OAuth and token management that keeps connections authenticated. We're also core contributors to the MCP spec — now an open standard under the Linux Foundation — and maintain the official Python and TypeScript SDKs.
We're looking for experienced software engineers to join Connectivity. The team spans San Francisco and London, and you'll work closely with partner teams across Anthropic to deliver impact across all of our products. You'll work on problems where reliability and enterprise trust are critical. This means OAuth at scale, admin controls that let IT govern what agents can do, and proxy infrastructure that is reliable by default.
What you'll do
You'll design, build, and operate the core connectivity infrastructure behind external tool calls — the MCP proxy and the OAuth and token management systems that keep connections authenticated. You'll own reliability for infrastructure that other teams and external partners depend on, including SLOs, on-call, incident response, and postmortems. You'll build the admin controls and permission models that let enterprise IT govern what agents can do, and contribute to the MCP standard and its ecosystem, including the spec and the official Python and TypeScript SDKs. You'll turn ambiguous outcomes into shipped, scoped work, dogfooding what you build in the products that consume it.
You'll independently scope complex, multi-month projects, drive cross-org alignment through ambiguous problem spaces, and make architectural decisions that shape how Anthropic builds and scales its products. You'll partner directly with research to productize cutting-edge capabilities, and will have lasting impact on the platform that hundreds of thousands of companies and internal/external engineers depend on every day.
You might be a good fit if you:
- Experience as a backend or platform engineer
- Familiarity with using AI tools as part of the software development process
- Take a product-focused approach to platform work and care about building solutions that are robust, scalable, and easy to use