About
Roman Di Domizio
Engineer, builder, and entrepreneur working at the edge of secure AI systems.
What I do
I'm an AI and cybersecurity engineer focused on secure infrastructure, LLM workflows, automation, and technical product development. I'm also a lifetime hockey player and athlete, and I stay active in Colorado through skiing, hiking, and biking. My goal is to build systems that are secure by default, observable in production, and useful to the people operating them.
Lockheed Martin Space
After 10 weeks as a Cyber Security Engineer Intern at Lockheed Martin Space, I transitioned into a full-time role as a Cyber Systems Security Engineer Associate. I now support cybersecurity, software assurance, accreditation, and AI-powered automation for satellite systems. Day to day, that means security-related work, software product assessments, STIG-based configuration review, and exploring where automation can reduce manual effort without adding risk.
I keep the scope of my Lockheed Martin work limited to what is public and non-sensitive, with no classified details, internal program names, or nonpublic launch information.
Building AI platforms
Before Lockheed, I worked as an AI Platform Engineer at WeOwnNet, where I built self-hosted AI infrastructure on Docker, Kubernetes, and DigitalOcean. The stack included PostgreSQL with PGVector, Infisical for secrets, OpenRouter for inference, MCP integrations, GitHub Actions for CI, and Braintrust for LLM observability.
I also researched and documented a migration path to vLLM and LiteLLM projected to cut inference costs significantly. It taught me as much about total cost of ownership as it did about model serving.
Product, agents, and leadership
Through The Verse, HackCU, and independent projects, I've led work on agentic systems, RAG pipelines, LLM simulation tools, AI product development, and full-stack applications. That range, from prototype demos to production infrastructure, taught me how to move between technical execution and product strategy quickly, and how to ship work that a team can actually maintain.
What's next
I'm turning this growing AI engineering skill set into useful products and businesses. Two ventures are currently in early development: an AI developer tool focused on making LLM-powered engineering faster and more reliable, and a structured AI consulting practice that helps teams adopt AI with real engineering discipline. Neither is announced yet.
Foundations
I hold a B.S. in Computer Science from the University of Colorado Boulder. While there, I also competed as an ACHA Division I hockey player for the CU Buffs. Playing at that level shaped how I approach teamwork, leadership, discipline, and competition. The same habits carry over to building in high-stakes technical environments.
Beyond the screen
When I'm not working, I'm usually on the ice, snowboarding, hiking, biking, or in the mountains around Boulder. The physical side keeps the technical side sane.