Researcher, Automated Red Teaming
Core
Building scalable, research-driven systems to continuously uncover failure modes in AI models and safeguards, translating findings into production improvements to reduce expected harm.
Role type
Senior IC research engineer (automated red teaming)
Builds
Automated pipelines for classifier jailbreak discovery, bio threat-development elicitation, and CoT monitoring evasion probing
Domain
AI safety, catastrophic risk mitigation, adversarial ML
Deliverable
production ML models
Required skills
applied research instincts, LLM and agent expertise, scalable automation engineering, software engineering fundamentals, threat modeling, experimental design
Preferred skills
adversarial ML experience, security research, abuse prevention systems, large-scale eval infrastructure
Technologies
LLMs, agents, automated pipelines
Responsibilities
Own research and technical direction for automated red teaming across catastrophic risk areas; partner with vertical risk teams to define threat models and prioritize targets; collaborate with Classifiers team to convert attacks into training data and evals; ensure ART outputs are operationally useful for stakeholders
Seniority
Senior, hands-on IC