Background · Role · Operating Style
I build AI-enabled military exercise products. CADE went from idea to execution in one week.
I work in the military training domain, designing and developing exercises for brigade units and below. My focus is practical: use AI to accelerate the work without handing over judgment, source truth, or risk control. Three live executions. An average of 19 participants per session. No simulation infrastructure required.
AI is the production engine. Not the design authority.
AI is fast. It is not reliable without a human in control of the inputs, the acceptance criteria, and the approval gate. I design the exercise logic, govern the source material, and decide what changes after live execution. AI drafts, checks, verifies, aligns, and revises inside those constraints.
What I build
CADE controller packages, operational orders, annexes, adjudication aids, evaluation tools, and the AI workflows that generate and govern them. Every product traces back to an approved source layer before it reaches execution.
Who I build for
Units that need realistic training products under time pressure. Controllers and trainers who need artifacts that work in the room without designer support. Leaders who need AI used responsibly, with review gates and traceable outputs.
AI Exercise Design and Operations Products Lead
That title describes the role I've built through CADE: leading the integration, governance, and applied use of AI to accelerate military exercise design and operational product development.
The work includes building AI tools when the mission requires them, integrating those tools into production workflows, advising leadership on AI use, and managing AI-related risk for the products I create.
Demonstrated Capabilities
What this looks like in practice
- Exercise design: Designed CADE from problem statement to live execution in one week. Three runs. Average 19 participants. No simulation infrastructure.
- Controller package production: Generated five-document controller packages from OPORD source truth using Claude, ChatGPT, and Gemini — trainer-reviewed before every execution.
- Operational product development: Produced OPORDs, annexes, op-boards, and supporting planning documents using AI-assisted drafting with human-governed approval at every gate.
- AI workflow governance: Separated AI production from design authority across every CADE artifact. Source-truth discipline enforced. No downstream product allowed to drift from the approved source layer.
- Evidence-driven iteration: Identified controller package fragmentation as a failure point in the first execution. Redesigned the artifact architecture before the second run.
Speed only matters if the product stays credible.
The goal is to produce usable products in days instead of weeks while preserving source truth, doctrinal alignment, human review, and execution value.
CADE is the clearest example. Three executions. Nineteen participants per session on average. One week from problem statement to execution-ready controller package. AI made that timeline possible. The training logic, adjudication model, controller discipline, and final approval remained human-led throughout.
CADE went from problem statement to live battalion staff exercise in one week and has run three times without simulation infrastructure.
Staff operated inside the structure. Controllers executed with delivered artifacts. Consequences created decision pressure. Learning was captured before the next turn.
Independent controller teams ran CADE without the designer present. The Controller Package was the only artifact they needed.
Explore the work
The CADE case study and AI operating decisions show how this works.