Intelligence Security Laboratories careers

Intelligence Security Laboratories is building secure infrastructure for transformative AI: implementation-driven R&D for high-security AI systems, based on a fail-safe cyber physical approach that is closer to avionics or reactor control than enterprise software. Our approach is described in this RAND report: rand.org/pubs/research_reports/RRA4827-1.html

A transformatively capable AI model is expected within a handful of years. If that model is not secure, it can be tampered with or stolen, and we cannot count on noticing this happening. We expect whoever builds it to reach for the most available off-the-shelf security, and we expect that security to fail.

A secure data center is only proven secure if someone owns the whole problem as an integrator. That is the role we fill. It means designing against the attack vectors nobody has thought of yet rather than the ones we can list, and it means the standard is no-fail rather than best effort.

We are a philanthropically funded 501(c)(3) organization based in Washington, DC, currently around ten people and growing quickly. We are tool-agnostic, and equivalent platforms and experience count. What matters more is depth across networks, operating systems, and security, and an interest in the problem itself.