Must have experience with hardware and software infrastructure requirements for effective LLM and ML development, to include GPU/TPU compute ecosystems, data center architecture, semiconductor supply chains, distributed training frameworks, and inference optimization techniques.
Must have experience with AI software development stacks, including familiarity with dominant frameworks (e.g., PyTorch, TensorFlow, JAX), model evaluation and benchmarking methodologies, and AI safety and alignment research.
Must have knowledge of current and near-term AI development trends, to include frontier model capabilities, multimodal systems, autonomous agents, dual-use AI risks, and the competitive global AI development landscape among state and non-state actors.
Must have experience with AI governance, policy, and international regulatory frameworks, including experience engaging with or advising government, academic, or industry stakeholders on AI policy development.
Must have knowledge of the current state and projected trajectory of AI capabilities relevant to arms control verification, treaty monitoring, and on-site inspection activities, including the use of AI in imagery analysis, data fusion, anomaly detection, and compliance assessment.
Must understand risks posed by adversary use of AI technologies to undermine arms control verification measures, spoof monitoring systems, or accelerate WMD development and delivery system programs.
Opportunities to leverage AI tools to enhance USG arms control implementation capabilities, improve training effectiveness, and support knowledge management within GT-AT and the Directorate.
Will provide the technical feasibility, verification challenges, and policy implications of potential arms control frameworks addressing AI development, deployment, or associated compute and hardware infrastructure.
Will assess the technical verifiability of proposed AI-related arms control measures, including limitations on compute thresholds, model capabilities, dataset size, or specific application domains such as autonomous weapons systems.
WHAT YOU’LL NEED TO SUCCEED
Education: Bachelor of Arts/Bachelor of Science
Experience: 8+ years of related AI experience and prior experience with national security, defense, or arms control contexts is highly desirable.
Security clearance level: Secret clearance and US citizenship required.