Staff+ Site Reliability Engineer, Safeguards ML Infra

EngineersOfAI; San Francisco, Northern (CA, KY); Hybrid; USD 180,000 - 240,000; Full time

Job summary

Anthropic is seeking a production-focused engineer for the Safeguards ML Infra team in San Francisco. You will own deployment and verification of safety safeguards across model launches, automate runbooks, and drive continuous validation to reduce manual work.

You will manage off-cycle classifier deployments, monitor configuration drift, and participate in on-call rotations to ensure reliable, secure operations of safety-critical systems.

Qualifications

• Experience in production change management at scale.
• Proven track record with deploy pipelines and canary analysis.
• Experience handling high-stakes releases and incident response.

Responsibilities

• Launch captain model releases: configure and verify safeguards for new models.
• Own off-cycle deployment of safety classifiers and post-deploy checks.
• Verify safeguards are live across all deployment platforms and deter drift.
• Automate deployment workflows and turn runbooks into tooling.
• Contribute to safety-oriented automation for continuous validation.
• Maintain a safeguards registry with provenance and deployment details.
• Participate in on-call rotations for incidents and time-sensitive launches.

Skills

Release management

Config management

Canary analysis

On-call experience

Python

Cloud platforms

Incident response

Tools

AWS

GCP

Job 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:

The Safeguards ML Infra team designs, builds, and operates the production infrastructure that powers Claude's safety systems. We own the critical backend services that ensure safety on the token generation path, and we own the operational work of getting those systems safely into production: standing up safeguards for every new model launch, and deploying new safety classifiers as they ship. Every frontier model release runs through this team – we configure, verify, and roll out safeguards across every platform Claude runs on (1P, AWS Bedrock, GCP Vertex, etc.), and we lead incident response when issues arise.

This role sits at the center of that operational work. You'll ensure safeguards are properly configured and deployed for model launches and own the off-cycle deployment of new safety classifiers — canarying changes, verifying that the right safeguards are provably live on the right models, and holding rollback authority when something looks wrong. Every launch should also shrink the checklist, and the manual verifications should evolve into a system that runs itself. You'll turn launch runbooks into tooling, hand-built checks into continuous validation, and one-off deploys into a repeatable pipeline.

We're looking for engineers with deep experience in production change management at scale — people who have owned deploy pipelines, config management systems, rollout safety, or launch readiness for systems under real production pressure. Familiarity with ML research or transformer architectures is not required — you will learn that on the job. What we prioritize is production judgment: a track record of shipping changes to critical systems safely, and of automating yourself out of the work you did last quarter.

What you'll do:

• Launch captain model releases: stand up, configure, and verify safeguards for every new model, and serve as the safeguards point of contact in the launch room during release windows.
• Own the off-cycle deployment of new safety classifiers as they ship from research — canarying rollouts, running post-deploy validations, and investigating discrepancies when something looks wrong.
• Verify that the right safeguards are provably live on the right models across every deployment platform (1P, AWS Bedrock, GCP Vertex, etc.), and detect and eliminate configuration drift between them.
• Automate yourself out of last quarter's work: turn launch runbooks into tooling, hand-built checks into continuous validation, and one-off deploys into a repeatable pipeline.
• Plan to use Claude aggressively to do this! And be a trailblazer that paves the path for safe agentic operations of safety-critical systems.
• Build and maintain a safeguards registry with full provenance — what is running in production, on which model, on which platform, and when and by whom it was deployed.
• Participate in on-call and operational-duty rotations covering service incidents, model provisioning, and time-sensitive research and safety launches.

You may be a good fit if you:

• Have owned production change management at scale — deploy pipelines, config management systems, canary analysis — and have strong opinions about what "verified" means.
• Have run high-stakes releases: served as a launch captain, incident commander, or release owner for systems where a bad deploy has real consequences, and are energized rather than drained by being in the critical path.
• Have meaningful on-call experience for production systems, including incident response and postmortem-driven improvements — and a track record of turning (and fixing!) postmortem action items into process and tooling changes.
• Have a desire to close the gap where nobody has yet raised their hand, even if it requires manually hand-holding processes until automation and tooling can be built.
• Have hands-on experience deploying and operating on cloud platforms (AWS, GCP) at scale.
• Are proficient in Python; experience with Rust is a plus but not required.