A practical, project-based course for experienced developers who want to use Claude AI and Claude Code in a disciplined engineering workflow. Turn requirements into testable plans, investigate an unfamiliar repository, implement focused changes, and critically review generated code.

Level: Advanced developer workflows. Estimated study time: 18 hours, including hands-on labs. Prerequisites: proficiency in one programming language, Git, automated testing, and command-line basics.

Course outline

1. Engineering with Claude (2 hours): task decomposition, context selection, acceptance criteria, and verification.
2. Repository onboarding (3 hours): map architecture, identify dependencies, and document project conventions.
3. Claude Code workflows (4 hours): investigate code, plan changes, implement a feature, and inspect diffs.
4. Debugging and test design (3 hours): reproduce a defect, build regression tests, and validate a repair.
5. Responsible automation (2 hours): permission boundaries, secrets hygiene, untrusted instructions, and human review.
6. Capstone (4 hours): deliver a tested feature and technical handoff in an existing application.

Learning outcomes

• Write precise engineering tasks with measurable acceptance checks.
• Evaluate AI-generated implementations instead of accepting them blindly.
• Produce a reviewable change set with tests and documentation.

Independent Eustad training; not affiliated with or endorsed by Anthropic. Tool access and any subscriptions or API charges are separate from tuition. Features may change.

Technical reference: Official Claude Code documentation.