Independent researcher in LLM-augmented systems engineering.

I explore whether a single developer can build non-trivial systems software by using LLM agents as bounded executors inside strict architectural contracts.

Sometimes I build games.
Sometimes the game becomes a stress test for an entire engineering methodology.

• LLM-augmented systems engineering
• Contract-driven architecture
• AI-agent execution boundaries
• Moddable simulation engines
• Documentation governance
• Native runtime / engine migration
• Long-horizon solo development methodology

Dual Frontier is a moddable simulation-engine project and a research artifact.

It tests whether a solo developer can coordinate complex software development through:

• locked architecture documents
• milestone briefs
• AI-agent execution sessions
• test-driven closure criteria
• governance registers
• risk / CAPA / audit tracking
• native kernel and runtime migration experiments

The game layer is the workload.
The engineering system is the experiment.

Human direction owner
        ↓
Architecture contracts
        ↓
LLM agent execution
        ↓
Tests, audits, closure protocol
        ↓
Measured artifact