Sean Lavigne builds execution authority and governance systems for AI agents - the layer between what an agent proposes and what it's allowed to do.
Former Nuclear Strike Advisor to the President. AI agent safety architect. Product leader who scaled federal platforms from 0 to national.
Agent Gate
Open-source policy enforcement layer that sits between AI agent reasoning and tool execution. Applies nuclear command and control principles, Permissive Action Links, to ensure every autonomous action is authorized, every destructive action is reversible, and the agent can never destroy its own safety net. Live-tested with Anthropic's Claude Code. Read the case study. View on Github.
eApp
I led the development of eApp from 0 to 1, the U.S. government’s cloud-based SaaS platform for background investigations, replacing e-QIP. I defined the roadmap, worked directly with agencies like the FBI and EOP, and applied analytics to reduce rework by 80% and cut form time from a week to 30 minutes. Read the case study.
WatchTower
I led the development of WatchTower, a case management and analytics tool that eliminated a 4-year backlog and delivered over $16M in savings by modernizing investigative workflows. Built as a field agent and scaled as product manager, it now supports thousands of federal users. Read the case study.
SENTRY
I designed and delivered SENTRY, a high-speed triage and analytics tool that automated military trainee identification across millions of records. By combining fuzzy logic and real-time data integration, I reduced a 4-hour task to 10 minutes. Read the case study.
FireSpec
I led the creation of FireSpec, a mobile-first iOS/iPadOS app that modernized fire safety inspections and uncovered a 10-year gap at rural schools across Oklahoma. It replaced paper workflows with digital reporting, mapping, and real-time dashboards. Read the case study.
Certifications
SAFe Agilist
AI-Empowered Product Owner
Product Manager
Responsible AI
with SAFe
Data Analytics Professional
Product Management