ai-ux

MagenticLite is an experimental agent that works across the browser and local file system

A shared language for designing and communicating AI experiences.

Split-pane canvas UI for OpenClaw agents — dashboards, kanban boards, charts, spreadsheets, and more, rendered by AI on the fly

AI Interactive — The interaction layer where intent flows freely between humans and AI. An open exploration of human-AI interaction patterns.

The 12-skill AI UX design engine for Claude Code & GitHub Copilot — purpose-built for designing UX for AI products, agents, and experiences.

AI Thinking Animation Generator

A self-contained design system for AI-assisted workflows and agentic interfaces — copilots, workflow automation, incident triage, research assistants, prompt orchestration, and human-in-the-loop dashboards. Built with semantic design tokens, dark/light themes, and an executable component contract for AI coding agents.

Final project for Product School's AI Product Management certification. Juno PM, an AI copilot that turns a forming P0 Slack thread into a ranked, source-cited shortlist a PM can defend: system prompt, build/buy call, RAG PRD, AI-native user flow, agent workflow spec with control panel, and a three-layer eval stack.

GestureBind is a cross-platform desktop app that maps hand gestures (via webcam) to system actions like hotkeys, app launches, or shell commands. Built with Python, OpenCV, MediaPipe (YOLOv8 planned), and PyQt5, it features YAML configs, profile management, system tray support, and on-screen feedback — all processed locally for privacy.

Most failed enterprise AI is a pattern problem, not a model problem. A Claude Skill that picks the interaction pattern and autonomy level for a service journey, with the organizational cost, trust mechanics, and metrics attached.

A research prototype in UI/UX: a mechanical bench for working with complex structures of meaning.

A structured, repeatable system for designing and prototyping AI-enabled product experiences.

State-of-the-art email productivity interface built with Next.js 16 & Ant Design 5. Features dynamic Kanban triage, real-time updates, and Gemini-integrated email intelligence.

Sentinel — embeddable inline AI oversight. Wraps the host AI's output in place so the human expert can validate, correct, and audit without leaving their workflow.

Real-time oversight of LLM agents — watch the agent execute, approve or edit every tool call, abort if it goes sideways. Third piece of the trilogy with Sentinel + Recourse.

A working UX systems prototype for agentic AI travel experiences, modeling explicit states, autonomy boundaries, confidence signals, and recovery patterns.

Design vocabulary for AI surfaces that are honest about uncertainty. Calibrated confidence, evidence anchors, the cross-hatch primitive — shared across the agentic-UX trilogy.

Production-tested UX patterns for AI products: streaming, RAG citations, token-cost transparency, quota fallback, evals, empty states — with real screens

Design experiments in code. Interactions I couldn't answer in Figma.

AI & UX Q&A Knowledge Base — 2,001 cards, zero-dependency scaffold for AI-assisted UX design and software engineering topics.