Course 03

Claude Code: Web / App Development — Claude AI training for developers

Claude Code: Web / App Development is a two-day course (7 hours per day) for software and full-stack developers, IT/DevOps teams and technical leads, and does require a programming background. It goes from CLI and IDE setup through reading, debugging and refactoring an existing codebase, into Anthropic API integration, building chatbot and RAG applications, and the security, cost-control and deployment practices those systems need in production.

AudienceSoftware and full-stack developers, IT/DevOps, technical leads

PrerequisitesProgramming background required

Duration2 days (7 hours per day)

LevelIntermediate to advanced

DeliveryOnline or onsite in Bangkok, in Thai

FormatIn-house for one organization

What does the course cover?

Claude Code CLI and IDE extension setup
Codebase understanding — reading, debugging and refactoring
Feature development and automated testing
Anthropic API integration in Python and Node.js
Building AI-powered web apps — chatbots and RAG
Security, cost control and deployment

Who is this course for?

Developers and technical leads who are either already building with Claude or evaluating whether to. A programming background is required — this is the one course of the three where that is a hard prerequisite.

It suits teams inheriting a codebase nobody currently understands as much as teams starting something new, since codebase comprehension is a large part of day one.

What can participants do afterwards?

Drive Claude Code from the CLI and from their IDE on a real repository
Use it to read into an unfamiliar codebase, then debug and refactor with tests
Integrate the Anthropic API into a Python or Node.js service
Build a working chatbot or retrieval-augmented generation application
Reason about token cost, prompt-injection risk and deployment before shipping

Why does the last module cover security and cost?

Because those are the two things that turn a working prototype into a production incident. Cost, because a loop that looked cheap in testing behaves differently at real volume; security, because an LLM application accepts untrusted input by design and prompt injection is not a theoretical concern.

Organizations putting these systems in front of customers usually pair the course with AI security and governance, which handles the same questions at the policy rather than the code level.

Claude Code CLI IDE Extension Anthropic API Python Node.js RAG Chatbot