About this role
Build and maintain the agent harness and evaluation infrastructure for NVIDIA's AI Safety & Security Engineering team, ensuring reproducible experiments and versioned tooling; collaborate with security research and evaluation engineers and work with agent frameworks such as NVIDIA NeMo. Responsibilities include harness development, building systems to run and reproduce experiments, owning environments and tooling, and partnering on researcher workflows.
Skills for this role
PythonContainerized environmentsCI/CDAgent frameworksLLM orchestrationML infrastructureEvaluation harnessesReproducibility practicesVersioned toolingNVIDIA NeMoSecurity toolingVulnerability researchInternal developer tools