Course Details
A 1-day course, 6 hours (10:00-17:00), run as a hands-on workshop throughout. Intermediate level; learners should already have HTML, CSS and JavaScript fundamentals and must bring their own laptop (Windows or macOS) with permission to install software. Learners take home the source code, a spec template set (design.md and spec.md), a prompt workflow, and their own application already deployed to production. The content focuses on workflow and principles, so it transfers to other AI tools as the technology changes.
Session 1: AI UI Design, MCP and Spec (10:00-12:00)
Section 1: AI-Assisted UI Design
- Using AI to draft screens and produce design alternatives quickly
- Evaluating AI-generated design: consistency, real usability and buildability
- Establishing a color system and component set so every screen stays coherent
- Lab: design your project screens with AI and pick the direction to build
Section 2: MCP, Bringing Design into the Editor
- What MCP is and how it lets an AI agent reach external tools and data
- Passing the design into the editor to generate starting code instead of from scratch
- Reviewing and adapting the generated code to fit the real project structure
- Lab: connect your design into the project and get your first screen scaffold
Section 3: Spec-Driven Development
- Why one-question-at-a-time prompting produces code that does not fit the codebase
- Writing design.md to describe the look and behaviour of the system for the AI to read first
- Writing spec.md to define scope, requirements and acceptance criteria per task
- Lab: write design.md and spec.md for your project and have the AI follow them
Session 2: Planning and Next.js Fullstack (13:00-15:15)
Section 4: Project and Task Planning with AI
- Having AI plan the project and order the work from the spec you wrote
- Breaking large work into pieces the AI can finish in one pass
- Sequencing tasks so development keeps moving without getting stuck
- Tracking progress and adjusting the plan when requirements change
Section 5: Next.js and TypeScript
- Next.js project structure and the split between server and browser work
- Routing, building APIs and fetching data for display
- Using TypeScript to reduce errors and give the AI a clearer view of the code
- Lab: scaffold a Next.js project with your first screens and API routes
Section 6: Databases with Prisma and Supabase
- Designing the data model and the relationships between tables
- Using Prisma to define the schema and manage structural changes
- Connecting Supabase and reading and writing data from the app safely
- Lab: build your project database and wire it into the screens you made
Section 7: Authentication and User Permissions
- Building signup and login with Better Auth
- Managing sessions and protecting pages that require login
- Defining user roles and checking permissions on the server side
- Lab: add login and restrict access in your own application
Session 3: AI Debugging, Agents and Production Deploy (15:15-17:00)
Section 8: Debugging and Code Review with AI
- Giving the AI enough context that it can actually find the cause
- Having AI review code and flag security and performance risks
- The limits of AI review and where a human still has to decide
- Delegating work to an AI agent in iterations and reviewing each round
Section 9: Deploying to Production and Wrap-up
- Preparing the project for deployment and the checks to run first
- Setting environment variables and storing secrets safely
- Connecting the production database and updating the app after deployment
- Workshop: deploy your app to production and plan how to apply this workflow