Teams at Lila Sciences

Lead security investigations for AI systems used in scientific discovery. Develop threat models, test defenses through adversarial assessments and technical analysis, define security requirements, evaluate evidence, and drive improvements with Security, Safety, and engineering teams. The role addresses dangerous misuse, model and data compromise, deployment risks, and safeguards for increasingly capable AI systems.

Build and maintain software, services, data pipelines, and model integrations for AI safety evaluations, red teaming, mitigation experiments, and scientific AI research. Translate researcher needs into reliable tools, preserve reproducible experiment records, and own systems from prototyping through deployment, testing, monitoring, documentation, and improvement. Collaborate with AI and software engineering teams on shared infrastructure.

Develop and implement safety mitigations for frontier scientific AI models, agents, tools, and laboratory-connected workflows. Translate risk assessments and policy requirements into safeguards spanning training, classifiers, monitoring, access controls, and runtime interventions. Design experiments, evaluate robustness and legitimate-use impacts, investigate failures and distribution shifts, and integrate mitigations with model training, inference, platform, and evaluation teams.