Hillsboro, Ohio, United States
Most AI conversations focus on models or applications. The harder problem appears once systems begin executing decisions inside real organizations. At that point the architectural question shifts from capability to control: where authority is validated, where admissibility is computed, and where execution becomes irreversible. My work focuses on the control plane of AI systems — designing architectures where autonomous agents operate inside explicit governance boundaries. This means systems where: • decisions cross defined commit boundaries • execution follows authority validation and constraint checks • actions remain observable, reversible, and audit-traceable • propagation risk is detected before state mutation Practically, this involves designing runtime structures that combine governance control planes, diagnostic layers, commit-boundary enforcement, and distributed agent coordination. As organizations move toward agentic workflows and automated decision systems, the central challenge is not just intelligence. It is governable execution at machine speed.
Design of governance, control planes, and decision boundaries for AI and automated systems, ensuring legitimacy, risk containment, and appropriate use under uncertainty.
Hands-on work with large language models (LLMs) and AI tools including ChatGPT and Claude, applying prompt engineering, prompt iteration, and systematic output evaluation to real-world analytical and communication problems. Practical experience with embeddings and retrieval-augmented generation (RAG) concepts, including analysis of model behavior, failure modes, alignment constraints, and retrieval quality. Comfortable working with databases and data analysis tools (SQL-based systems, Excel, Python notebooks, BI-style dashboards) to extract insights from structured and unstructured data. Strong analytical skills translating complex technical systems into decision-ready insights and clear narratives for non-technical stakeholders. Produce written artifacts and explanations that combine data storytelling with content creation, delivering technically sound, audience-appropriate outputs that prioritize accuracy, trust, and appropriate use over hype.