The opportunity
The Security Risk team is responsible for how Anthropic identifies, prioritizes, and drives treatment of its most important security risks. We are rebuilding risk management to operate as an engineering function through automation and AI-native platforms to enable decision making.
What you'll do
Take ownership of Anthropic’s most complex security risk problems and drive: them end to end with minimal oversight, turning ambiguous signals into a defensible view of severity and likelihood and seeing them through escalation, treatment decisions, and remediation
Build the systems that make risk measurable and let risk work scale,: including quantification tooling, automated intake and triage, and the observability that partner teams use to understand their own risk posture
Work alongside Security Engineering as a calibrated technical peer who: pressure tests architectures and treatment plans, translates findings into prioritized remediation roadmaps, and makes the investment case for what to fix now, what to accept and track, and what to defer
Mentor engineers and risk practitioners across Security and the broader: engineering organization, and help build a risk engineering culture in which teams own their risks and our team provides the visibility and judgment that supports the
Security risk engineering Work across the breadth of Anthropic’s security: landscape, spanning areas like identity and secrets management, developer security and supply chain, infrastructure security, and secure frameworks, and build the context needed to reason about risk credibly in each
Go deep where a risk demands it, understanding how these systems are built: and how they fail so that assessments and treatment plans reflect engineering reality rather than abstraction
What they're looking for
- Owned a named security risk and driven it from discovery through remediation across multiple teams
- Briefed executives on risk decisions and defended accept, remediate, or: transfer recommendations under challenge
- Built security automation, detection, or risk platforms adopted across an engineering organization
- Shipped LLM or agent-powered tooling and workflows that automate security or risk activities
- A security engineering, detection engineering, or offensive security: background with a risk-based prioritization mindset
- Built or operated a quantified security risk program (FAIR-style: decomposition, Monte Carlo simulation, loss exceedance analysis) whose outputs changed real resource decisions
- Enough familiarity with SOC 2, ISO 27001, or FedRAMP to know what compliance does and does not buy you
