Labs
Public builds for AI workflow governance, knowledge systems, career evidence, and experiments. The applied AI systems are part of the portfolio; the rest remain here for anyone who finds them useful.
Applied AI systems
A context-safe delegation model for AI subagents, built for Codex and Claude. One router prevents agent sprawl, routes work to the fewest useful capabilities, passes compact state packets, preserves ownership boundaries, and quality-gates handoffs so supervision, trigger tuning, and least-privilege delegation stay visible.
A private cross-tool AI memory for one person: keep decisions, preferences, and project context available across Claude, Codex, and Cursor, without restating the same background in every new task. It keeps project-specific guidance discoverable too, so local ways of working do not disappear into a generic global rule.
A team-safe second brain for shared work: decisions, operating context, and source-grounded knowledge remain available across Claude, Codex, and Cursor, while the guidance that belongs to one project stays with that project. Human approval keeps new shared memory trustworthy.
A copy-improvement tool, not an AI verdict machine. It asks whether a draft says anything before it asks how the draft sounds: a paragraph carrying no name, no number with its mechanism, no bounded event, and no decision is filler that no line edit will save. Then it corrects the generic, over-processed, and AI-shaped patterns that survive. It returns corrected copy across portfolio pages, professional writing, summaries, resumes, and cover letters — never a score, never an authorship verdict.
Career tools
Collect and rank public employer postings so you can spend less time on stale, duplicated, closed, or low-signal roles. It filters by fit, salary, location, posting age, and source health, then exports a review list. Runs entirely in your browser.
Ground every resume and cover-letter claim in real evidence, fit notes, and proof narratives, then hand off to DOCX. A voice gate checks each bullet for a real detail before checking its wording, and reports a bullet that has none as an evidence gap rather than rewording it into something smoother. It doubles as an applied AI workflow-governance build — evidence-grounded generation with human review gates.
Practice and interview from your own evidence, not canned answers. It turns prepared background notes into a Story Log, keeps relevant talking points handy during the conversation, and never speaks, types, or invents answers for you. Windows desktop app; bring your own OpenAI key.
Projects
Public experiments in interaction design, generative visuals, and playable browser experiences.
A collection of full-screen animated hero banners: portfolio signal cards, AI decision gates, and filament-network fields, each with parallax and pointer response.
A high-velocity browser arcade shooter with weapon pickups, boss waves, touch controls, keyboard movement, and a particle-heavy starfield.
An 8-bit radial arcade shooter with ship capture, a solar-system tour, drifting asteroids, and a synthesized techno soundtrack.
Portfolio return path
The main portfolio remains focused on PMO, portfolio governance, executive operations, AI workflow governance, evidence discipline, and delivery follow-through.
Connect
For PMO, portfolio governance, AI workflow governance, and executive operations evaluation, start with the portfolio, experience, and proof pages.