Machine Learning Engineer - Secure AI Lab

About the role

The Secure AI Lab within the SEI’s AI Division focuses on improving the security and robustness of AI systems. The lab conducts and applies cutting-edge research to protect AI systems from adversaries who aim to manipulate the system to learn, do, or reveal something it isn’t supposed to.

Responsibilities

• Identifying and investigating emerging AI and AI-adjacent technologies.
• Defining and refining processes, practices, and tools for working with AI.
• Designing and building well-engineered prototypes of AI systems.
• Transitioning and providing guidance on AI capabilities to government sponsors.
• Building Machine Learning Models and Systems: Work with machine learning frameworks such as TensorFlow, PyTorch, Torch, and Caffe and modern programming languages including Python, C/C++, and Java.
• Experimenting with modern and emerging machine learning frameworks, methods, and algorithms in application domains that include computer vision, natural language processing, planning and scheduling, robot control, and engineering safe, trusted, and reliable machine learning systems.
• Testing and evaluating systems for performance and security.
• Mentoring and contributing to improving the overall technical capabilities of the Division by mentoring and teaching others, participating in design (software and otherwise) sessions, and sharing insights and wisdom across the SEI.

Requirements

• A bachelor’s degree in computer science, statistics, machine learning, electrical engineering, or related discipline with eight (8) years of experience; OR MS in the same fields with one (1) year of experience; OR PhD in a relevant discipline with two (2) years of experience.
• Willingness to work onsite 5 days per week at SEI offices in Pittsburgh, PA or Arlington, VA.
• Subject to a background investigation and must be able to obtain and maintain an active Department of War security clearance.
• Willing to travel up to 25% of the time to locations outside of your home location.