I design the architecture of AI conversation: the state machines, evaluation rubrics, prompt frameworks, and behavioral guardrails that make models coherent, safe, and actually useful. Psychology training. A decade of operational leadership. AI product management certification for the strategic layer.
Designed a 12-phase conversational state machine with anti-loop and anti-dependency guardrails for a sensitive emotional wellness product. Wrote evaluation rubrics that score model behavior across safety, tone, and pacing.
Turned user and stakeholder input into conversation flows and AI task structures for a career guidance platform. Built onboarding quality criteria, memory accuracy benchmarks, and standards for interaction consistency.
Defined agent roles, interaction flows, and evaluation criteria for a coordinated multi-agent system built for entrepreneurs. Found UX and logic gaps across agent outputs. Validated the whole thing in a live keynote demo.
State machines, branching logic, anti-loop rules, and multi-turn flow design for high-stakes AI interactions.
Structured rubrics, failure mode mapping, behavioral benchmarking, and reproducible feedback cycles. Grounded in UX research.
Writing prompt frameworks that control model tone, pacing, accuracy, and boundary behavior — consistently, at scale.
Safety guardrails, dependency prevention, and quality thresholds for AI that operates in emotionally sensitive territory.
Most AI systems don't break at the level of language. They break underneath it, in the layer where timing, restraint, and decision-making live. When to engage. How to pace. What to allow. What to withhold. Those are the things that determine whether a system feels coherent or falls apart under its own output. That's the layer I work in.
My background is in psychology and UX, which keeps me grounded in how people actually think, feel, and respond — not just what they say. Operations experience means I know how to build systems that hold up in the real world, not just on paper. AI product management training ties it together and helps me work across the full lifecycle, from early research through to production.
Before AI and UX, I spent nearly a decade in operations across technology organizations. The most recent stretch was four years as Director and Manager of Business Operations at a global cybersecurity firm during a period of 250% growth across multiple countries and U.S. states. I worked directly with executive leadership on budgets, resource allocation, and compliance. Led SOC II audits. Built workforce analytics dashboards. Cut hiring time-to-fill by 30% through process redesign.
That work gave me a ground-level view of how enterprise teams function, where processes quietly fail, and what it actually takes to build systems that outlast their first quarter. That's the lens I bring to every AI project.
