Mark Myshatyn is a program manager for the U.S. Department of Energy’s Genesis Mission, which brings together artificial intelligence, supercomputers, scientific instruments, and specialized research data. He works on making advanced AI useful for scientific discovery and national security without compromising the security, accountability, and operational controls those fields require.
His background spans cybersecurity, cloud infrastructure, compliance, and applied AI in sensitive government environments. As enterprise AI architect at Los Alamos National Laboratory, he helped shape AI deployment across scientific and operational workflows. In 2023, he coauthored research on evaluating foundation models for nuclear nonproliferation, addressing how emerging systems could be assessed for security-sensitive analytical tasks.
Scientific AI under real-world constraints
- Agent-driven scientific discovery: Myshatyn describes an inertial confinement fusion workflow in which an agent reviews scientific literature, proposes a fusion-capsule design, and tests it using hydrodynamic and thermodynamic simulations. The system joins language-model reasoning to established physics tools and high-performance computing, allowing scientific hypotheses to face computational scrutiny.
- Security-conscious AI infrastructure: His work connects advanced models to institutional knowledge, specialized simulation software, and infrastructure such as Los Alamos’s Venado supercomputer. He identifies explainability, isolation, and governance as practical deployment requirements for systems operating around controlled information and classified environments.
- Federal deployment readiness: Myshatyn argues that vendors serving government laboratories must plan for continuous monitoring, software bills of materials, procurement requirements, federal authorization frameworks, and deployment inside restricted environments. He also emphasizes workforce readiness, human-AI teaming, and collaboration among national laboratories, universities, and commercial AI developers. His AI Engineer presentation on government AI agents connects these requirements to concrete scientific applications.
By 2026, Myshatyn had brought that combination of scientific computing and security architecture to the Genesis Mission, extending his focus from laboratory-scale deployment to a broader national effort linking AI, research infrastructure, energy innovation, and scientific discovery.
