AI Safety: Agentic Alignment Enforcement Analyst

Candidate; Boston (MA); Vor Ort; USD 245.000 - 285.000; Vollzeit

Benefits dieser Stelle

Generous vacation and parental leave

Flexible working hours

Equity donation matching

Zusammenfassung

Anthropic is seeking an Enforcement Analyst focused on Agentic Alignment. You will identify and investigate harmful and unexpected model behavior, and make enforcement decisions in collaboration with Policy, Threat Intel, Engineering, Research and Transparency teams.

This role involves exposure to explicit content and on-call responsibilities across Policy and Enforcement. You will work with Claude and other tools to summarize transcripts, analyze code, and build lightweight tooling to scale

Qualifikationen

• Ability to read and understand code in Python, JavaScript, or shell scripts.
• Background in security, cyber threat intelligence, or a similar technical field.
• Experience in trust and safety, enforcement, investigations or a related field.
• Structured investigations and evidence-based findings.
• Able to present complex findings to non-technical audiences.

Aufgaben

• Review and investigate potential instances of Agentic Misalignment.
• Triage cases by severity and urgency and share through appropriate channels.
• Operationalize review pipelines for agentic cases and scale workflows.
• Make enforcement decisions under the Usage Policy and document reasoning.
• Assess cyber-related agentic activity and determine real-world harm potential.
• Write clear, well-evidenced case write-ups for diverse audiences.
• Identify patterns across cases and feed findings back to policy and teams.
• Maintain rigorous documentation of decisions and precedents.
• Collaborate with Threat Intelligence, Research, Engineering and Policy.
• Use Claude and other AI systems to accelerate investigations.
• Contribute to regular reports on agentic harm trends.

Kenntnisse

Code reading

Python

JavaScript

Shell scripting

Security background

Enforcement investigations

Clear writing

Ausbildung

Bachelor's degree or equivalent

Tools

Claude AI

Analytics tools

Jobbeschreibung

Anthropic is seeking an Enforcement Analyst focused on Agentic Alignment. You will identify and investigate harmful and unexpected model behavior, and make enforcement decisions in collaboration with Policy, Threat Intel, Engineering, Research and Transparency teams.

This role involves exposure to explicit content and on-call responsibilities across Policy and Enforcement. You will work with Claude and other tools to summarize transcripts, analyze code, and build lightweight tooling to scale