AI Safety & Security Researcher - Innovate AI Systems

RAND Corporation; Washington (District of Columbia); Hybrid; USD 132,000 - 339,000; Full time

Job summary

RAND’s Center on AI, Security, and Technology (CAST) seeks researchers to advance AI safety foundations. The role targets deep experts in theoretical CS, cryptography, formal methods, PLs, security, or ML systems, with independence to formulate and execute challenging projects.

Researchers will conduct original work spanning theory, systems building, empirical evaluation, and analysis, communicate results through publications and policy-facing outputs, and collaborate across distributed teams

Qualifications

• Demonstrated ability to conduct independent, original technical research.

Responsibilities

• Conduct original research on AI safety at the intersection with security.
• Lead projects from problem formulation through analysis, implementation, experimentation, and publication.
• Communicate results through publications, technical reports, research artifacts, and policy-facing products for both technical and non-technical audiences.
• Work in a collaborative, geographically distributed environment.

Skills

Independent research

Theoretical computer science

Cryptography

Formal methods

Programming languages

Python

Rust

Go

C/C++

Systems security

Machine learning systems

Leadership

Communication

Collaboration

Education

PhD in Computer Science, Electrical Engineering, Mathematics, Physics

Master’s degree with 3+ years relevant experience

Bachelor’s degree with 5+ years relevant experience

Tools

Python

Rust

Go

C/C++

Job description

RAND’s Center on AI, Security, and Technology (CAST) seeks researchers to advance AI safety foundations. The role targets deep experts in theoretical CS, cryptography, formal methods, PLs, security, or ML systems, with independence to formulate and execute challenging projects.

Researchers will conduct original work spanning theory, systems building, empirical evaluation, and analysis, communicate results through publications and policy-facing outputs, and collaborate across distributed teams