"AI/ML Vulnerability Analysis Intern"

What We Do:

The SEI helps advance software engineering principles and practices and serves as a national resource in software engineering, computer security, and process improvement. The SEI works closely with defense and government organizations, industry, and academia to continually improve software-intensive systems. Our core purpose is to help organizations improve software engineering capabilities and develop or acquire the right software, defect free, within budget and on time, every time.

Position Summary:

The SEI Vulnerability Analysis Team's mission is to reduce the societal harm from vulnerable information systems and related processes. A rapidly expanding mission area is timely AI/ML threat analysis supporting AI security of the Defense Industrial Base and U.S. National Security Systems. The Vulnerability Analysis Team Intern will contribute to a variety of core functions in this expanded mission area including:

• Vulnerability analysis of specific AI/ML software systems, the implications, and mitigation recommendations;
• Strategic-level trend analysis of emerging AI/ML software system vulnerabilities and recommendations;
• Threat actor intent, capability, and opportunities to actively exploit legitimate AI/ML software systems; and
• Analysis of tools that leverage AI/ML in malicious ways with intended use by threat actors

Requirements:

• Enrolled in a degree granting program
• Able to work an internship during the summer
• Able to work onsite to the SEI Pittsburgh office
• You will be subject to a background check and must be eligible to work in the United States without Visa sponsorship
• Willingness to explore various locations to support the SEI's overall mission, specifically within the SEI and CMU community

Duties:

• Enable vulnerability analysis of AI/ML software systems including mitigation recommendations
• Enable trend analysis of emerging AI/ML software system vulnerabilities and recommendations
• Enable analysis of threat actor intent, capability, and opportunities to exploit legitimate AI/ML systems

Knowledge, Skills and Abilities:

• Ability to critically think and problem solve
• Basic understanding of cybersecurity and software development concepts
• Solid understanding of AI/ML
• Basic proficiency with a programming language (preferably Python)
• Familiarity with common AI/ML software packages and tools (e.g. Pytorch, TensorFlow, ART)
• Ability to collaborate internally with professional engineers and researchers
• Willingness to learn new technologies with cross-functional teams
• Potential to analyze code and system architectures to identify vulnerabilities
• Skills in communicating technical matters clearly and effectively to non-technical stakeholders
• Ability to communicate externally with government, academia, and/or industry stakeholders

Desired Experience:

• Experience with programming languages (preferably Python)
• Experience with training AI/ML models
• Experience with adversarial machine learning techniques
• Experience with red-teaming cyber and/or AI systems
• Basic understanding of cybersecurity concepts
• Experience with clearly documenting and communicating findings

Job Function Breakdown:

• AI Security and Development: 80%
• Documentation of Results: 20%

TOTAL = 100%

This position is based in Pittsburgh and is 5 days on-site.