The work

Contractors develop realistic, unclassified SIGINT scenarios and reference materials, then review AI-generated intelligence assessments for accuracy, relevance, consistency, and analytical quality. Day-to-day tasks may include identifying reasoning errors, checking whether operational workflows are represented credibly, and documenting where an output departs from established intelligence methodology.

Evaluation responsibilities

The work also involves creating and refining rubrics for model evaluation. Reviewers may need to distinguish a merely plausible response from one supported by the scenario, assess treatment of uncertainty and gaps, and provide structured feedback that project teams can use to improve model performance. All examples and discussion should remain unclassified and suitable for customer-facing AI evaluation work.

What the platform screens for

The AI interview is likely to verify specific government or military SIGINT experience, depth in collection or analysis workflows, and judgment when evaluating intelligence products. Background with the NSA, the US intelligence community, or allied government intelligence agencies is preferred. Expertise in COMINT, ELINT, signals analysis, or cyber intelligence operations is also preferred, as is experience with intelligence reporting, training, quality review, AI evaluation, model testing, or rubric development.

Logistics

This is a remote contractor role with flexible, project-based or task-based commitments. No typical weekly hours are stated, and the listing does not specify whether all collaboration is asynchronous. Pay band: $50–100/hr