Curated AI safety and governance jobs

Lead research on frontier AI risk modeling across cyber, CBRN, and loss of control, setting methodological standards that inform safety cases, evaluations, mitigations, and AI governance.

Technical Program Manager for Cohere’s public sector and defence AI deployments, with substantial responsibility for governance, compliance, risk assessments, and responsible AI adoption in sensitive government contexts.

Legal counsel role focused on EU AI and privacy regulation for enterprise AI products, including compliance programs, risk assessments, and trustworthy design.

Senior research engineering role focused on building evaluation methods, benchmarks, datasets, and infrastructure for measuring frontier LLM capabilities; relevant because it directly concerns model evaluations for advanced AI systems.

Product management role bridging Cohere's safety research and North product, translating model evaluations and red-teaming findings into safety features, guardrails, and evaluation frameworks.

Senior research role focused on creating next-generation evaluation methods, benchmarks, and infrastructure to measure LLM progress and model capabilities.

Forward Deployed Engineer role building and deploying LLM-powered agentic workflows for enterprise customers, with explicit responsibility for evaluation frameworks, safety, observability, and auditability of AI systems.

Senior technical staff role on Cohere’s Safety for Agents team focused on data generation, post-training algorithms, and evaluation methods to improve safety and trustworthiness of LLMs and agents.

Part-time remote contractor role annotating, auditing, and red-teaming LLM outputs to improve model safety, policy alignment, and prevention of unsafe or adversarial outputs.

Technical program manager for Cohere’s public sector and defence AI deployments, focused on delivery, governance checkpoints, regulatory compliance, and responsible AI considerations.

Technical role at METR focused on executing and scaling AI evaluations for capabilities, risks, mitigations, autonomy, and alignment.

Senior programs role at FAR.AI focused on designing and curating AI safety events, convenings, and field-building programming across the AI safety ecosystem.

Remote hourly contract role analyzing AI and LLM security scenarios, probing frontier models for vulnerabilities, misuse, and adversarial threats; relevant because it directly involves AI system safety, red-teaming, and misuse prevention.

Research role focused on frontier model evaluations and environments to improve agent capabilities while steering models toward safe AGI/ASI, in collaboration with safety/alignment partners.

A full-time, in-person research fellowship in Singapore focused on technical AI safety and governance, with projects aimed at frontier AI safety practice and policy translation.

Lead EMEA assurance programs and compliance controls for Scale, including AI governance compliance intersections, certifications, audits, and customer assurance across public sector and commercial business.

Editorial role on OpenAI’s Safety Systems team focused on producing and improving public-facing transparency materials about technical safety work, including evaluations, safeguards, red teaming, and deployment decisions for frontier models.

Deputy Head of Programme (Security and Technology) Location Geneva Closing date 25 June 2026 Contract Temporary Position Title: Deputy Head of Programme (Security and Technology) Starting date: Negotiable, as soon as possible Occupational rate: 100% (Full-time) Duration: 364 days, with the possibility of extension Remuneration: P-4 Application Background The United Nations Institute for Disarmament Research (UNIDIR) is a voluntarily funded, autonomous institute within the United Nations. One of the few policy institutes worldwide focusing on disarmament, UNIDIR generates knowledge and promotes dialogue and action on disarmament and security. ...

Technical AI Safety Project Sprint BlueDot Impact Remote, Global This course teaches how to make a meaningful contribution to AI safety research or engineering through a structured 30-hour project sprint. Includes weekly mentorship from an AI safety expert to scope, refine, and execute your project with rapid feedback. Publishes projects on the course website as a public signal of your technical AI safety skills to collaborators and employers. Provides collaborative check-ins with peers to receive feedback and build your AI safety portfolio. BlueDot Impact is a nonprofit that runs courses which aim to help participants develop the knowledge, ...