Marsel Bait

I'm Marsel Bait, a product designer in Tokyo.

I design products for two kinds of users: Humans + AI Agents.

see my work

Selected work

Enjin.world

Redesigning a bilingual recruitment platform

Design ownership on the full platform redesign (Web + Mobile). Three sub-brands, one system. Sole designer on the 2023 revamp after four years on the product.

Open case study
Enjin mobile login screen, navigate your future, held in hand
Nankai Trough Hazard Map

A bilingual hazard map on cited government data

Japan's official Nankai Trough earthquake and tsunami forecasts, on one bilingual map. Search any address for the projected shaking and tsunami height.

Open case study
Nankai Trough seismic intensity map of Japan, color graded by shaking strength
manabiAI

A current, credible face for a course in AI and n8n

Landing site for a course in AI and n8n workflow automation. Visual design, layout, and system for a young, AI-fluent audience.

Open case study
manabiAI landing page shown in a floating browser frame on a dark backdrop
Miraidukuri

Design for a Tokyo social enterprise

Tokyo social enterprise spanning tourism, regional revitalization, health, and the circular economy. Design across tour booking, brand identity, and print.

Open case study
Miraidukuri site hero, timeless wisdom future vision, over a Japanese kaiseki table

Tools for AI agents

My view is that AI agents are becoming a real part of how customers interact with products. I have spent my career designing for people, and I think this next chapter asks us to design for agents as well. That is what MCP servers are to me: interfaces agents call directly, built with the same care as any UI. Nine shipped on npm, around 4,800 downloads a month.

Also: Nankai Trough, Rakuten, Xendit, tabedata, pdf-it, MCP Apps UI. More on GitHub and npm.