
Research Engineer, AI Monitoring
AI
Anthropic is seeking a Research Engineer for AI Monitoring to build, deploy, and monitor classifiers that detect unsafe agent actions in real time. The role combines research on high-recall, high-precision classifiers with production engineering to ship them into critical paths. Requires strong prioritization, code review, and communication skills, with a sincere interest in safe AI.
What you'll do
- Design, train, and evaluate classifiers that detect unsafe agent actions in real time
- Deploy, tune and operate classifiers in critical paths, owning accuracy, reliability, latency, cost, and failure behaviour
- Build datasets, labelling pipelines, and evaluations to measure and improve classifiers and harm taxonomies
What you need
Must-have
Machine LearningClassifier DevelopmentLarge Language ModelsPythonProduction SystemsCode ReviewExperiment DesignData Labeling PipelinesEvaluation
Nice-to-have
LLM InferenceModel InternalsFine-tuningSecuritySandboxingDetection and ResponseAI ControlScalable Oversight
Soft skills
OrganisationPrioritisationCommunicationCritical ReviewInterest in AI Safety
AI SafetyResearch EngineerMachine LearningClassifiersLLMProduction EngineeringHybridLondon