Transluce
AI Behavior Engineer (Forward Deployed)
Why it's interesting
$310K–$500K at a nonprofit is the plot twist of the week — Jacob Steinhardt and Sarah Schwettmann's eval lab pays frontier-lab money for an FDE whose 'customers' are the EU AI Office and civil-society partners, with privileged-access evaluation exercises run inside frontier labs. Deployment work where the deployment target is AI oversight itself.
The role
Work directly with policymakers, civil society partners, and frontier labs to rapidly answer why AI systems act the way they do and when they fail. Execute on Transluce's government-evaluator contracts (including building harmful-manipulation evals with the EU AI Office), design privileged-access oversight exercises with frontier labs, and adapt behavioral-evaluation pipelines to partner contexts (mental health, persuasion, evaluation awareness). Based in San Francisco.
Applications go directly to the company. We curate; we don't broker.