About this role
This role focuses on building an adversarial function to rigorously test and critique agentic reasoning systems, identifying flaws and vulnerabilities in automated judgments. The researcher will design experiments to probe system weaknesses, such as poisoning and injection, and evaluate the impact of AI outputs on human decision-making.
Skills for this role
Adversarial MLSecurity ResearchMulti-agent SystemsReinforcement LearningGame TheoryForecastingCalibrationCausal InferenceExperimental DesignStatistics