Luigui Calderin Portfolio
Client/Company
None
Project
Personal
Role
UI/UX Designer

This portfolio is, above all, my professional introduction to companies and freelance projects. But it's also a personal project that allowed me to explore different aspects of my professional profile, with a special focus on UI design.

The foundation of my portfolio

AI is a powerful tool for boosting productivity, and although this portfolio could have been built 100% with prompts, I'm an advocate for personal judgment when you have the tools and knowledge to apply it. The initial sketches were generated with Figma Make, providing it with the colors, typography, and style of my personal brand. From there, everything was built from scratch.

40% IA
The initial visual structure was generated with Figma Make.
60% manual
Every element was adjusted, rearranged, and redefined with personal criteria, to reflect my identity as a designer.
Figma Make screenshot with the prompts used to build the portfolio.
Figma Make screenshot with the prompts used to build the portfolio.
Design System

Building a Design System for a personal portfolio might seem excessive, but when it serves both as a professional introduction and a demonstration of knowledge, the investment makes complete sense. It was the ideal opportunity to showcase expertise in color scales, states, tokens, variables, components, and documentation.

Image of Foundations defined for the portfolio.
Image of Foundations defined for the portfolio.
Imagen de componentes tipo Button.
Image of Button-type components.
Setting up Variable Tokens

Primitive and semantic tokens were defined for colors and measurements, establishing a solid and scalable foundation for the system.

Image of primitive color tokens.
Image of primitive color tokens.
Image of semantic color tokens.
Image of semantic color tokens.

A good practice when working with variables is to keep existing styles, but link them directly to those variables to ensure consistency across the system.

Image showing the link between color variables and color styles.
Image showing the link between color variables and color styles.
Dark Mode

Not all portfolios offer the option to toggle between light and dark mode. Defining both modes in this project was a decision driven by accessibility and visual comfort for the user.

Image of the portfolio's Home in light mode and dark mode.
Image of the portfolio's Home in light mode and dark mode.
Component documentation

Since this is a system designed exclusively for this portfolio, I chose to document only the Button component. To do so, I used Claude to generate the base documentation from specific instructions and references, focusing my work on reviewing, refining, and correcting the output.

Image of the Button component documentation.
Image of the Button component documentation.

95% of the design system was built manually, with custom criteria and definitions tailored to this project. The remaining 5% corresponds to documentation generated with AI, guided by references from design systems I've previously worked on.

Moving to Webflow

With basic knowledge of HTML and CSS and experience with no-code/low-code tools, I decided to manually transfer the entire design from Figma to Webflow.

Through the Figma to Webflow plugin, I exported the project variables —typography, colors, and measurements— directly into Webflow. The entire site was built manually, but for the parts that required code, I used Claude through the Webflow MCP. This made it possible to implement the light/dark mode toggle, as well as animations and microinteractions throughout the site.

Image of my personal portfolio workspace in Webflow.
Image of my personal portfolio workspace in Webflow.
50% IA
Implementation of light/dark mode, hover effects, microinteractions, and text adjustments.
50% manual
The entire site structure: divs, containers, headings, and pages.

The conclusion of this project reinforces a clear idea: AI is a tool, not a substitute for knowledge. It helps increase productivity and streamline repetitive processes, but it should never be indispensable for completing a job.

Using AI to design without understanding design, or to code without understanding code, is like learning to drive only in automatic: it works fine until the day you're handed a manual car. The same applies to AI — anyone who relies on it completely will be left without tools the day they can no longer use it.

Other projects
Descriptive image from project
Bancolombia
La Cuenta del Mar