Staff / Principal Research Engineer, AI Safety Mitigations
Core
Develop and test safeguards to reduce societal risks from closed-loop scientific AI models, agents, and laboratory-connected workflows.
Role type
Staff/Principal Research Engineer (AI Safety)
Builds
Safety systems, mitigation designs, and experiments for scientific superintelligence models.
Domain
AI Safety / Scientific Superintelligence
Deliverable
production ML models
Required skills
Safety intervention development, frontier model training/post-training, model-level classifiers, inference safeguards, abuse monitoring, adversarial testing, experimental design, robustness assessment, research engineering, cross-team collaboration.
Preferred skills
Scientific models, tool-using agents, multi-step workflows, policy-to-control translation.
Technologies
Model training pipelines, inference infrastructure, evaluation frameworks, automated safety systems.
Responsibilities
Translate risk assessments into concrete mitigation designs; develop and implement safeguards across training, inference, and runtime; build experiments measuring risk reduction and robustness; investigate regressions and misuse patterns; integrate mitigations into model development and serving systems.
Seniority
Staff/Principal, hands-on IC with strategic scope.
