Designing in Code: Figma, MCP and Production Handoff
For designers and front-end-minded product people who already know Figma and want to move designs into real code. You will learn to connect design files to AI coding tools and ship changes to a production codebase.
8 lessons · 2 hr 7 min, plus 1 optional (25 min more with the optional lessons)
Start here: short framing of designing in code with AI tools and why design systems keep the result consistent.
Explains how SaaS product designers can design interactive interfaces directly in code using AI-assisted tools, covering workflows for new and existing products, the role of design systems and Figma, and when the approach suits a product team.
Adds a practitioner's view of designing in Cursor while the design system stays the source of truth.
An interview with design engineer Karl Koch of DuckDuckGo, who explains how he designs in code with Cursor instead of Figma and how that shortens the time to ship features. It covers importing Figma designs, using a design system as the source of truth, and testing and shipping changes.
Shows live demos of moving a design from Figma into code using GitHub and Copilot, the full handoff loop.
A Config 2025 session where GitHub's Head of Design and a Staff Research Engineer show live demos of moving work from design to code using GitHub, Figma and GitHub Copilot. It also previews directions from GitHub Next.
Adds hands-on setup of the Figma desktop MCP server, the technical base the next videos rely on.
A step-by-step Figma tutorial on setting up the desktop MCP server and connecting it to Cursor, VS Code and Claude Code. It then shows how to use the server to pass context from Figma design files into agentic development tools so code stays aligned with the design.
Compares Figma Agent and Claude with MCP on one ticket, showing what MCP output still needs fixing by hand.
The video compares Figma Agent and Claude with Figma MCP by giving both the same real product ticket and design-system requirements, then reviewing how well each understood the task, respected the system and made UX decisions. It is aimed at designers who want to judge AI design tools on real work.
Optional: adds an end-to-end build from Claude Design through Cursor to deployment for extra practice.
A creator builds a party game called Face Card in four hours, moving from a speech-to-text brainstorm to a design in Claude Design, a handoff to Cursor and Claude Code, and a Vercel deployment. It covers steering AI for UI motion and debugging the database and API keys, and it is aimed at designers who want to build with AI tools.
AI Design Professor | Nehmat Gereige · 7 min · Intermediate
Introduces Code Connect mapping in a production codebase, preparing you to work in a real repo.
A developer walkthrough of setting up Figma Code Connect with Claude from a production codebase, mapping a Material UI button to its Material 3 design and checking the result in Storybook and Figma. It is aimed at developers, design engineers and designers who want to understand how code and Figma libraries connect.
AI Design Professor | Nehmat Gereige · 24 min · Intermediate
Applies Claude Code to a production React repo with Storybook, ending in a pull request for review.
A walkthrough for designers on editing a production React codebase with Claude Code, using GitHub, Storybook and Material UI components, and submitting a pull request for developer review. It is aimed at UX, product and design systems designers who want to contribute to code.
For designers new to AI who want to understand its role in design work and try beginner-friendly tools. You will be able to judge whether a tool fits a task and turn a prompt into a working prototype.
For intermediate designers who want to take a UI from layout grid to working code. You will plan interfaces, build interactive components, document them for developers and use AI to generate screens from a design system.
For intermediate UX and product designers who want to use Claude and Figma's agent in a consistent way. You will learn to build repeatable design workflows and choose AI tools sensibly for your own process.
For designers comfortable with a little JavaScript who want to automate repeat design work. You will be able to write small Figma plugin scripts, build tools with AI coding assistants, and judge when a modular framework fits.