I design with production in mind.

I don't design for static mockups. I shape products with the system, constraints, and implementation in mind, so they can move smoothly from design into a working product.

AI expands the search space. Judgment narrows it. Systems make the process repeatable. Real users make it better.

Stage 01

Explore broadly

Every project starts with research, ideation, and a shared brief: the problem, the audience, the outcome, the constraints. I explore that brief through several prototyping setups to open a wide, comparable range of directions.

With visual references Without references With a design system Without one Constrained prompts Open prompts

Tools

Ideation

ChatGPT

Research

Perplexity

Rapid prototyping

Lovable

Base44 Figma Make

Stage 02

Curate and sharpen

I read the directions as evidence, not candidates for a winner. I save the strong parts as fragments, flows, interactions, sections, and visual ideas, including good ideas from prototypes that fail as whole products. That library sends me back to research with sharper questions and a clearer visual direction.

Stage 03

Model the ecosystem

Before refining screens, I map the full user journey and what the product needs in order to work: flows, states, data, dependencies, backend needs, and edge cases. A product is not a mockup, it is an ecosystem. Here I decide what I build directly, what I prepare for a developer, and what needs a structured handoff.

Stage 04

Equip the project's operating system

Once the journey and scope are clear, I move into a focused production environment. MCP connections give it live project context: design files, code, browser state, and other relevant systems. Skills turn recurring methods, standards, and checks into reusable behavior.

Every project-specific skill carries explicit rules: what triggers it, what context it needs, when it asks for approval, when it consults another skill, and which decisions get a second and third quality check.

Discover capabilities Connect context through MCP Reuse skills Create project skills Define triggers, approvals, and checks

Tools

Build environment

Claude

VS Code

GitHub

Stage 05

Build and refine

This is the hands-on production phase. After comparing the directions, I choose a path and build the final version. I combine the strongest decisions from the earlier exploration, then refine the flows, interactions, visual language, and implementation until the product reaches the quality bar.

Tools

Production

Claude

Figma

Stage 06

Finish and ship

The final route depends on what the product and its users need.

Fidelity first

When visual precision is central, I move the working product into Figma through MCP. I complete the detailed visual work, refine the design system, then return the product to implementation and verify it against the approved design.

Design refinement → Design system → Implementation and QA

Design

Figma

Feedback first

When users are already waiting, I set a clear quality threshold, publish the working product, and improve it from real use.

Define good enough → Publish to users

Implementation

Claude

Build Publish Learn Improve

Both routes end in the same loop. A high-fidelity product still needs real feedback, and a fast release still needs a quality standard. The goal is to put the product into use and keep improving.