PLUS is a tutoring platform that connects human tutors with AI-driven software to boost learning gains for middle school students from historically underserved communities. The platform currently supports over 13 schools, 3000 students, 500 tutors, completing more than 10k hours of tutoring every month.
As the Product & Design Systems Designer, I built an AI-powered prototyping agent using Figma tokens to generate system-aligned UI, enabling faster validation across design, product, and engineering. I led the scaling of the design system by standardizing tokens and component sizing, and by migrating legacy work into a unified Web App Specs file to improve consistency and developer handoff.
Design and prototyping work at PLUS involves repeated context re-entry and tool-switching as ideas evolve from low to higher fidelity. Designers often redo setup work, re-explain system constraints, and reconstruct context when moving between tools or collaborators.
At the same time, PMs and developers lack a fast, reliable way to visualize ideas early. Existing tools either require deep design knowledge or produce outputs that are inconsistent and difficult to evaluate.
Instead of focusing on a specific tool, we decided to plan the foundation of what we wanted to build.
PLUS UNO, an AI-assisted prototyping workflow that combines design-system context, a prototyping workspace, and shared review infrastructure. Its role was not to replace every design tool. It was to help teams move from rough ideas to system-aligned, evaluable prototypes much faster.
It works as an interactive source of truth. Instead of static specs, the team had a place where they could inspect components, states, and behavior in a format that was much more usable during design and implementation.
To bring AI guidance directly to where designers work, I integrated a live LLM backend into Storybook. By embedding the AI directly into our component library, designers and developers can pull up usage guidelines, get technical breakdowns of the specific screen they are viewing, and navigate the design system using natural language.
A centralized platform to host all prototypes for easy viewing so the team can review, compare, revisit, and reuse outputs.
A structured framework that tells the AI how to behave depending on the task — whether that's consulting on layout or rapidly prototyping a new idea.
I created onboarding materials and workflow documentation for PLUS UNO, enabling 10+ designers to more efficiently adopt AI-assisted prototyping workflows. Supported scalable team adoption by improving workflow clarity, onboarding consistency, and cross-functional collaboration.
Our initial goal was to generate production-ready code directly from prototypes, but after multiple iterations and dev team reviews, the AI-generated code—especially using a legacy HTML/CSS/jQuery-based stack—was not usable. The coding agent struggled with legacy patterns, and the generated code was difficult to validate, review, and maintain.
We structured the design system into modular markdown files — components, tokens, layouts — router files would direct AI to read only the relevant context for a given task, rather than processing everything at once.
Our hope was to package the tool up into an NPM package so that it could be become tool-agnostic - especially utilizing it in Figma Make. However, it didn't work, and we decided to stick with Cursor, since it had the strongest capabilities.
As PLUS UNO stabilizes, our focus shifts toward maturity and expansion. We aim to solidify the agent’s place in the design process and explore new opportunities for impact.
.png)
