Course

AI for Code & Development

Ship a real, deployed full-stack app built primarily with AI agents. Cursor and Claude Code, MCP, sub-agents, debugging, and the production lifecycle.

Length
~16 hours + capstone project
Structure
12 modules · 46 lessons
Certificate
Verifiable on completion

Coming soon

This course is still in development. The structure below is a preview of what will be available — modules and lesson titles may change as we build it.

Back to all coursesStart with Foundations
1

The Agentic Coding Landscape

  • Vibe coding vs. vibe engineering vs. agentic engineering
  • Tool landscape: Cursor, Claude Code, Codex CLI, Copilot, Windsurf, v0, Lovable
  • When to use which tool
  • Token economics and model selection (Opus vs. Sonnet vs. Haiku)
2

Context Engineering for Code

  • What context engineering actually is
  • CLAUDE.md / .cursorrules / AGENT.md patterns
  • Structuring repos for AI-friendliness
  • The art of the brief: specs that produce good code
3

Working with Cursor

  • Cursor fundamentals: chat, composer, agents
  • Rules, indexing, and project setup
  • Tab completion vs. agent mode
  • Multi-file edits
4

Working with Claude Code

  • Claude Code in the terminal: a different mental model
  • Slash commands and CLAUDE.md
  • Plan mode, edit mode, and approval flow
  • Headless and CI usage
5

Specs, Plans, and Checkpoints

  • Writing specs that produce good code
  • Planning multi-file changes
  • Checkpoint-driven development
  • Breaking work into agent-sized chunks
6

MCP (Model Context Protocol)

  • What MCP is and why it matters
  • Configuring MCP servers (filesystem, git, browser, database)
  • Building a simple MCP server
  • Multimodal MCP: Figma → code, screenshots → code
7

Agents, Sub-agents, and Skills

  • When to use sub-agents vs. main context
  • Building Skills (consistent methodologies)
  • Hooks: PreToolUse, PostToolUse, Stop
  • Long-horizon autonomous tasks
8

Code Review & Verification

  • Never trusting AI output blindly
  • The test/lint/run loop
  • AI as code reviewer (reviewing your own AI code)
  • Detecting subtle bugs and hallucinations in code
9

Debugging with AI

  • Explaining stack traces and errors
  • Reproducing bugs with AI
  • Performance debugging
  • When to stop letting AI debug
10

Production Lifecycle

  • Tests as the AI's safety net
  • CI/CD with AI
  • Security scanning and AI
  • GitHub PR autonomy
  • Vibe coding → production
11

Where AI Breaks Down

  • Codebases too big, contexts too long
  • Tasks too vague
  • Domain-specific knowledge gaps
  • When to write it yourself
12

Capstone

Ship a real, deployed full-stack app built primarily with AI agents.

  • Capstone Project