Shipping with AI

Read this end to end before Day 1. It will save you frustration, help you arrive ready, and put your mind in the right place.

• A note from the Instructor

  I want to say one thing clearly before we start. You are not behind. You are not slow. You are not someone who "doesn't get computers.

• What this course is, in one paragraph

  About Shipping with AI

• The week-by-week, in plain English

  An outline of what each week contains

• What we need from you

  Some things to setup before the class starts

• Frequently asked questions

  Some common questions about the course

• How to set yourself up for success in this course

  Some advice to follow

Theme: Build a thing. Put it on the internet. Feel what shipping feels like. Why this week exists: Most dropouts happen because students go weeks without producing anything they can show another human being. We end Week 1 with a live URL. Everything else in the course is downstream of that emotional win.

• Day 1 — The Reframe (live, 1 hour)

  The Vibe Coding Loop: Describe → Generate → Run → Read → Refine.

• Day 2 — Prompting for Code (live, 1 hour)

  The anatomy of a good prompt: context, specification, constraints, examples. Bad prompts vs good prompts (workshop with side-by-side comparisons). How to ask the AI to explain code it just wrote. Reading exercise: the AI's code from Day 1, explained line by line. Lab assignment: Improve yesterday's page. Add three sections. Make it responsive. Read every line and write a comment above it in your own words.

• Day 3 — Git, GitHub, and the Internet (live, 1 hour) — SHIP DAY

  What Git and GitHub actually are, in plain English. Setting up a GitHub account and your first repository (we use the web interface, not yet the command line — that comes Week 2). Deploying to Netlify via drag-and-drop. Your live URL. Friday Ship: Everybody pastes their live link in the cohort channel. The whole class visits each other's sites.

Theme: Move from the browser to the terminal. Meet the tools real developers use. Why this week exists: Browser AI chat is training wheels. The terminal is where AI tooling gets powerful — agents that can read your whole project, edit multiple files at once, run commands, and complete tasks autonomously. By the end of this week, students will be doing things that would have taken them weeks of traditional study.

• Day 1 — Demystifying the Terminal (live, 1 hour)

  What the terminal is. Why developers love it. Why it intimidates beginners. The five commands that get you 90% of the way: cd, ls (or dir), mkdir, pwd, and one more thing each platform needs. Installing Node.js (the engine many AI tools need). Installing a VS Code AI extension — Continue.dev or Cline, both free and work with free Google API keys. Lab assignment: Use the terminal to create a new project folder. Open it in VS Code. Use the AI extension to build a 'joke of the day' page that shows a different joke every refresh.

• Day 2 — Gemini CLI: Your First Terminal Agent (live, 1.5 hours)

  Why Google's Gemini CLI matters for our cohort: it is free, generous (1,000 requests per day on a personal Google account), open source, and runs in your terminal. Installation walkthrough. Authentication with a Google account (free tier covers nearly everything students will do in this course). First commands: asking it to read a file, summarise a project, suggest changes. Using Gemini CLI to refactor yesterday's joke page — making it cleaner without writing a line of code yourself. The GEMINI.md file: how to give the agent persistent context about your project. Lab assignment: Use Gemini CLI to add a feature to your Week 1 landing page entirely from the terminal. Document the prompts you used.

• Day 3 — Reading Code Like a Detective (live, 1 hour) — SHIP DAY

  The most important skill in vibe coding: knowing whether the AI gave you good code or bad code. Red flags in AI-generated code: hardcoded values where they should not be, files that grow forever, the same logic copied three times, made-up library names. How to ask 'is this good?' — getting a second opinion from a different AI than the one that wrote it. Workshop: instructor shares three buggy AI outputs. Students find the problems. Friday Ship: Each student pushes a v2 of their site to GitHub, with a written 'code review' of their own code in the repo's README.

Theme: Build something that does more than display text. Learn what makes a project feel like a real product, not a school exercise. Why this week exists: Up to now students have built single-file projects. This week we go multi-file, introduce real product thinking, and bring in design. This is the week where students stop feeling like beginners.

• Day 1 — Claude Code: The Senior Developer in Your Terminal (live, 1.5 hours)

  What makes Claude Code different from Gemini CLI: better at large, complex changes; understands your whole project; explains its reasoning. The honest cost conversation: Claude Code requires Claude Pro at about $20/month. Whether that is worth it depends on what you are building. We will demonstrate live so students can decide. Installation walkthrough (native installer, no Node.js required). The CLAUDE.md file — like GEMINI.md, but for Claude Code. Plan Mode: how Claude Code can plan before it writes, so you do not get blindsided. Live demo: instructor takes a feature request and builds it across 4 files in 10 minutes. Lab assignment: Students who can afford it set up Claude Pro and Claude Code. Students who cannot continue with Gemini CLI. Both groups start their Week 3 project (see Day 2).

• Day 2 — Choose Your Project (live, 1 hour)

  Each student picks one project from a curated menu. Examples: A to-do list with categories and a streak tracker. A simple expense logger with charts. A landing page with a working contact form that emails you. A flashcard study app that loads cards from a JSON file. A 'Naija news aggregator' that pulls headlines (uses an API — instructor provides scaffolding). Each project requires multiple files (HTML, CSS, JavaScript in separate files at minimum). Each student writes a one-page spec for their project before writing any prompt. This is the most underrated coding skill in the world. Lab assignment: Use your CLI agent (Gemini CLI or Claude Code) to build the first version end-to-end.

• Day 3 — Design With Claude, and Why Design Matters (live, 1.5 hours) — SHIP DAY

  Why design separates 'looks like a student project' from 'looks like a real product.' The minimum design vocabulary: spacing, hierarchy, contrast, alignment, restraint. Using Claude (web or terminal) for design prompting specifically — asking it to critique your design, suggest a colour palette, suggest typography, find what is wrong with a screenshot. Tools we touch: Tailwind CSS (instructor demos how to ask the AI to use it), shadcn/ui patterns for those going further. Live workshop: each student shares a screenshot of their project in chat. The class and instructor critique it together. Then they prompt Claude to fix the top three issues. Friday Ship: v1 of the Week 3 project, deployed, with a before-and-after design screenshot in the README.

Theme: Build something that is not a website. Learn a little Python the vibe way. Graduate. Why this week exists: Most students think 'coding' only means websites by this point. This week we show them what else is possible. We build a small command-line tool, touch a little Python, and integrate everything they have learned into a portfolio-grade final project.

• Day 1 — Hello Python, The Painless Way (live, 1.5 hours)

  Why Python is worth knowing even for someone aiming to be a web developer. Installing Python (or using an online sandbox for students whose laptops resist). Building a simple Python script with the help of Gemini CLI or Claude Code: a tool that takes a folder of images and renames them in bulk, or a tool that reads a CSV and prints a summary, or a tool that scrapes a single web page. Reading the Python out loud, line by line. Noting the differences from JavaScript. Lab assignment: Build a Python script that does one useful thing for you personally. Examples: cleans up your downloads folder by file type, generates a weekly summary from a CSV, sends a daily email reminder (with a real free email API).

• Day 2 — Final Project: The Capstone (live, 1 hour)

  Students propose their final project. It must integrate at least two things they have learned: a website plus a backend script, a CLI tool with a small web dashboard, or a multi-page app that calls an API. Instructor approves the scope (most beginners pick projects too big — this session prevents heartbreak). We introduce Vercel and Render for hosting things bigger than a static site. Pair-up: each student gets a 'ship buddy' they check in with daily this week. Lab assignment: Ship a working v1 by Thursday night. Spend Friday polishing.

• Day 3 — Demo Day (live, 2 hours) — GRADUATION

  Each student demos their final project in 3 minutes: 1) What is it? 2) Who is it for? 3) Show it working. 4) One thing you learned about yourself by building this. Class votes on three awards (decided per cohort): Most Useful, Most Beautiful, Most Surprising Use of AI. The Pediforte team gives a short closing talk on what comes after this course — paths into freelancing, paths into deeper study, paths into building your own products. Certificate of Completion issued.