<div class="content-intro"><h2><strong>.
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.
</p></div><h2 id="about-the-role">About the role</h2> <p>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.
</p> <p>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.
</p> <h2 id="what-youll-do">What you'll do</h2> <p>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 w.