Archive-style profile
Technical Profile Record
Archive Report Visual Concept
Personal Profile Record
Record Date: 2026
Writing that connects architecture theory with production decisions: how AI-native systems are structured, governed, funded, evaluated, and improved in operation.
Each record opens the complete article on its original publication site. Presentation slides, videos, and handbook reference pages are outside this index.
Frames token usage as an operating choice between efficiency, balance, and aggressive modes, then connects each mode to business conditions and staffing responsibilities.
Published by reopt architecture
Explains how execution traces, evaluation data, and learning signals turn operational history into an improvement loop that remains independent from any model.
Published by reopt architecture
Connects token unit cost and margin to three ways an AI product can sustain its operating loop: self-sufficiency, investment, or ecosystem leverage.
Published by reopt architecture
Describes how systems can preserve control of their learning loop while keeping human direction responsible for goals, limits, and final judgment.
Published by reopt architecture
Maps the Assess, Define, Loop, Execute, and Evaluate methodology onto repository instructions, commands, sub-agents, and skills in Claude Code.
Published by reopt architecture
Shows how a production SaaS closed governance gaps with approval records, audit logs, MCP tracing, rate limits, agent constraints, and operational visibility.
Published by reopt architecture
Demonstrates how a design-governance harness can translate architectural principles into manifests, contracts, audit rules, approval gates, and measurable product structure.
Published by reopt architecture
Archive report is the visual design concept only. This is Eric Han's first-party personal profile; the professional records are factual, not a leaked, investigative, or third-party dossier.