DeepLearning.AI and JetBrains · on DeepLearning.AI
Teaches writing markdown specs and project constitutions so coding agents build what you intend, then planning, implementing and validating features, applying the workflow to legacy code, and packaging it as a reusable agent skill. For anyone who has outgrown loosely directed vibe coding and wants more predictable results from agents.
Aspiring builders, career switchers and working developers. This is a beginner-level course in Coding with AI, running about ~1.3 hours (15 video lessons).
Free to audit. Videos are free with a DeepLearning.AI account; the graded quiz and accomplishment require DeepLearning.AI Pro ($25/month billed annually or $30/month billed monthly). Price seen 2026-11-01.
Moves from vibe coding to disciplined agentic coding with Claude Code, covering specs, test-driven development and context engineering, and ends with building a Tic-Tac-Toe game in Node.js. For beginners who learn best with interactive lessons and want a certificate; Codecademy suggests its Learn Prompt Engineering course first.
Shows how agentic coding tools break down complex apps across multiple files, with hands-on work in Windsurf debugging and refactoring JavaScript, updating a large legacy codebase, and building a Wikipedia data analysis app. For developers with basic coding and prompting skills who want to see how an AI IDE agent uses tools and context.
Andrew Ng walks complete non-programmers through describing an app idea in plain language and having AI build a working browser-based web app, then customizing it and reusing the approach on other projects such as a ping pong game. The gentlest on-ramp to vibe coding for anyone who has never written code.