Curated roles
Browse roles focused on reducing risks from advanced AI systems, including alignment research, safety engineering, evaluations and related operations.
343 active roles found.
Research Scientist focused on AI evaluation, language model robustness, red teaming, and alignment research for an automated evaluation platform for LLMs.
This role invites individuals to engage in research on AI model threats, participate in evaluations, and influence AI policy initiatives.
Software engineering internship building a reliability platform for LLM applications, with explicit work on evaluation models, dynamic testing, red-teaming, observability, and robustification.
The role focuses on building a platform for training and evaluating interpretable AI systems, emphasizing safety and reliability.
Research role focused on AI security, cyber capability evaluations, and mitigating misuse and dangerous agent capabilities for frontier AI systems.
Research engineer building evaluation and security testing systems for frontier AI models, including capability assessment, evasion testing, and risk mitigation.
Senior ML research role focused on developing models, experiments, and evaluation frameworks for AI safety problems, with explicit alignment and frontier-model safety relevance.
The role involves developing and evaluating probabilistic inference methods to support safe-by-design AI systems.
PhD research role focused on AI safety, evaluating large language models and frontier agentic systems to detect deceptive behavior and understand novel behaviors for safer AI development.
Senior research role focused on AI safety research, frontier-model red teaming, and evaluations in high-risk domains.
Leads Faculty's AI safety research team, overseeing research on safe language models and safety-critical systems, including evaluations and red-teaming in high-risk domains.
Scientist/Senior Scientist role focused on AI safety strategies, evaluations, and safeguards for scientific superintelligence models integrated with physical labs.
Senior developer role building an open-source security evaluation platform for AI agents, including scoring, auditable evaluation schemas, policy enforcement infrastructure, and threat modeling.
Full-stack software engineer building evaluation tooling and monitoring infrastructure for frontier AGI safety research, with direct support for scheming detection and mitigation work.
Backend software engineer on Apollo Research’s research team building evaluation tooling, LLM monitoring/control infrastructure, and other internal systems supporting frontier AGI safety research.
The role involves designing methods to ensure AI models align with goals, developing monitoring techniques, researching control mechanisms, and conducting red-team simulations.
Research scientist role focused on frontier AI risk evaluations, including designing evaluation harnesses and datasets, testing dangerous capabilities, and communicating results to policymakers.
The role involves running evaluations on frontier AI systems to assess risks, including pre-deployment evaluations and analyzing model behavior.
Research scientist/engineer role focused on studying scheming, misaligned preferences, and evaluation techniques for frontier AI systems.
The role involves developing methods to mitigate false information in large language models, creating evaluation protocols, and conducting research in LLM security and interpretability.