I build trusted product systems out of complex work.
I've spent 13+ years designing complex B2B systems where clarity, trust, and workflow fluency matter. Before AI became my default toolkit, I built design teams, design systems, feedback loops, and product workflows that helped organizations move faster without losing quality.
Now I operate through an AI-native stack: ChatGPT, Claude, Claude Code, Codex, GitHub workflows, custom skills, plugins, routines, Lovable, TanStack, Notion, Obsidian, Figma Make, Figma MCP, and live public-data prototypes. The transformation muscle is the same. The leverage is very different.
Same transformation muscle. More leverage.
- 01
At Optera, I led the pre-AI version of workflow transformation: design systems, feedback loops, research practice, supplier workflows, automation guardrails, and data-dense climate-tech product surfaces.
- 02
At Helm, I turned that practice AI-native: Claude Code skills, recurring audits, product principles, prototype loops, source-of-truth docs, and review gates that make AI work legible to design and engineering.
- 03
In FieldRules, I built the agentic version: a seven-layer run loop, a 2,265-page artifact system, a living skills registry, synthetic and adversarial evals, design-system automation, and coherence checks across Jira, Notion, GitHub, Supabase, Vercel, Slack, and Storybook.
- 04
In recent experiments, I build the whole loop myself: public-source data pipes, working interfaces, schema and evidence layers, provenance-aware product language, motion captures, and deployable artifacts.
What I'm best at
Turning ambiguous, high-stakes work into trusted product systems. Design systems, research practices, feedback loops, and the connective tissue between product, engineering, and customers. I work where complexity meets credibility.
How I lead
Quietly, precisely, and with high standards. I build small teams that ship durable patterns rather than one-off features. I mentor through routines and shared quality bars. I name the hard tradeoffs early so the team can decide together.
How I work with AI
AI is the environment I operate in, not a feature I bolt on. ChatGPT, Claude Chat, Claude Cowork, Claude Code, Codex, Lovable, GitHub PR workflows, TanStack Start, Notion, Obsidian, Figma Make, and Figma MCP run as one operating layer. Custom skills, Claude Code plugins, GitHub-sourced skills, and recurring routines do the repeatable work. Provenance, evals, and human review gates keep it trustworthy.
Why this matters now
Most teams are trying to add AI to a workflow that was already broken. The opportunity is to redesign the workflow itself — and to build the playbooks, guardrails, and rituals that let a team operate AI-native without losing judgment.
“product designer and a product thinker in the same person”
Michael Koenig
Product judgment
“one of the most creative and collaborative leaders”
Jenny Jones
Creative leadership
“benchmark for excellence that I still reference”
Hilary Rallo
Craft standard
Looking for senior design leadership.
Open to AI-native design leadership, design program management, product design, and workflow transformation roles.