What you will learn
- Distinguish autocomplete, chat, and agent modes, and choose the level of assistance that fits the task.
- Prepare a clear task, relevant repository context, and measurable acceptance criteria to support a good result.
- Break the agent's work into small, traceable steps while retaining control over decisions and changes.
- Do not treat a convincing AI explanation as proof: verify behavior through tests, documentation, and runtime results.
- Control access boundaries, sensitive information, and external effects when giving an agent context and tools.
This course is in development
Enrollment is not open yet. You can preview the curriculum below. Access and pricing will be confirmed when the course launches.
View the first available courseCourse curriculum
5 modules 20 lessons
- Differences between code completion, IDE/CLI agents, and cloud agents
- The difference between a vibe coding prototype and verified production code
- Task size, model limitations, and human oversight
- Time, token, and cost budgets; measuring output quality