Daniel
Clark
that ship world-leading AI products
I lead experience design across automation, low-code, and agent platforms, building mission-critical infrastructure that lets millions safely delegate work to autonomous systems with enterprise security and governance.
About
I've spent my career in enterprise software designing automation platforms used by millions — and the last two years making them the runtime for AI agents. Before that, I helped make Carbon IBM's open-source design system. I build what I design: MCP servers, agentic Figma plugins, open-source tools, and the hero playing above this page.
Things I built
shipped, open source, or running right nowOpen-source MCP browser — agents read authenticated pages on your machine, for a fraction of the tokens.
GitHub →The hero above — real GitHub events from the Claude Code, MCP and Anthropic ecosystem, replayed on canvas.
Running now ↑Hermes agent — answers questions about my work, routes to case studies, requests deep dives.
Open the agent →Autonomous publishing pipeline — agent essays go live with no approval step. Kill switch is a Worker secret.
Read →Selected work
artwork generated live in codeBuilding, developing and sustaining happy, healthy design teams, and the cross-functional partnerships around them, at scale.
Read the case study →Bring the experts to the problem. Then bring the agents.
Read the case study →Oversight surfaces for autonomous work — plans you can read, actions you can audit, failures you can repair.
Read the case study →The visual automation canvas admins and builders use to compose triggers, logic and actions without writing code.
Read the case study →MuleSoft Composer, Flow Integration and IDP: admins connect any system and any document from inside Flow, and agents act on both.
Read the case study →Led product and front-end design for Watson Analytics on Carbon — one open-source design language across IBM's product portfolio.
Read the case study →The first team in the Salesforce ecosystem to integrate SLDS 2, and an agentic Figma plugin that answers in SLDS 2 code.
Read the case study →An open-source MCP browser: agents read the pages behind your login, on your machine, for a fraction of the tokens and none of your keys.
Read the case study →