AI Dev
An AI coding agent will happily write more code than you can read. This course is about staying in front of that — using Cursor as a collaborator you direct and verify, not an autopilot you hope for the best with.
Five lessons, built around one rule: treat AI output as a first draft you guide and verify. Everything else — tool choice, prompting, context, rules, review — is machinery for making that rule practical on real codebases.
What you will be able to do
- Pick the right mode for a task — Tab, Ask, Plan, Agent, Background Agent, Bugbot — instead of defaulting to one
- Write prompts with a verification path built in, so the agent can check its own work
- Engineer the context an agent sees: what to include, what to exclude, and why more is not better
- Encode your team's standards once as
.cursor/rules/*.mdcand have them injected only when relevant - Review AI-generated diffs for the failure modes AI actually produces — hallucinated APIs, insecure defaults, silently removed safeguards
- Answer senior-level interview questions on AI-assisted engineering with a defensible position
The shape of the course
| Lesson | The question it answers |
|---|---|
| 1 · The collaboration contract | What am I responsible for, and which tool fits this task? |
| 2 · Prompting and context | What does the agent need to see, and how do I ask? |
| 3 · Rules | How do I stop repeating my standards in every chat? |
| 4 · Reviewing AI code | What do I hunt for before this diff reaches main? |
| 5 · Interview readiness | How do I defend all of this to a senior interviewer? |
The tooling will keep changing. The judgement — knowing what to delegate, what to verify, and what never to hand over — is the part that transfers.