Course Overview
AI coding agents can now plan, edit, test, and review software inside a development environment. This two-day workshop teaches you how to direct that work, keep it grounded in a real user need, and decide whether the result is ready to ship.
It's for experienced software developers. Through five practical projects with Claude Code or Codex, you'll move from small individual builds to an unfamiliar agentic codebase and a team AI product. Along the way, you'll work with live data, persistent project instructions, reusable capabilities, parallel workflows, external tools, and structured review.
The workshop combines product thinking with hands-on engineering. You'll spend less time typing every line and more time framing the problem, setting useful constraints, inspecting the agent's work, testing assumptions, and improving what you build.
Lesson Outcomes
By the end of this course, you will be able to:
Frame a useful product before building it. Identify the user, clarify the problem, control scope, and turn an idea into a plan an agent can act on.
Direct an agent through a real codebase. Provide durable context and constraints, manage long-running work, and choose when to prompt, plan, intervene, or rewind.
Connect software to the wider world. Work with live data, AI models, external tools, and reusable agent capabilities while handling configuration and secrets responsibly.
Collaborate safely with people and agents. Divide work, use branches and worktrees, review pull requests, and set clear goals and limits for parallel or cloud-based agents.
Judge and improve the result. Reproduce problems, test important assumptions, use evidence to prioritise changes, and polish a working product for its intended audience.
Curriculum
Setup and foundations
Prepare Claude Code or Codex, GitHub, and the course tools. We'll look at how coding agents work inside a development environment, what changes when implementation becomes faster, and why the developer remains responsible for the outcome.
Project 1: From problem to working product
Turn a short product brief into a focused application, review the agent's plan, and publish the result.
Practise managing context and recording instructions that should persist across a project.
Project 2: Building with live data
Explore a real data source, decide who it could help, and build a small product around it.
Learn how to inspect unfamiliar inputs, recover from an unhelpful direction, and keep the build aligned with its purpose.
Project 3: A deliberately small shared app
Scope and deploy a useful app for a specific group, gather feedback, and turn that feedback into clear engineering work.
Use reusable agent capabilities and parallel workflows to make and review improvements safely.
Project 4: Working in an agentic codebase
Trace how an existing AI application works, including its model, tools, and external services.
Reproduce a problem, investigate it with coding and browser tools, implement a fix, and review the change through a pull request.
Project 5: Team AI product
Define a user, desired outcome, and key assumption, then build one complete slice of an AI-backed product.
Extend the project with one deeper agent-engineering technique, improve the end-to-end experience, and present evidence of what worked.
Presentations and conclusion
Teams demonstrate their products, explain the problem and engineering choices, and reflect on the evidence, trade-offs, and next steps.
Details
Dates and Times: 10-11 Sep 2026 or 24-25 Sep 2026 (2 full-day sessions, 9:00 to 17:00).
Location: CT Hub 2, Lavender.
Requirements: Bring your own laptop. Modern Mac, Windows, or Linux laptops are supported. Participants should be comfortable using a terminal, Git, GitHub, and at least one AI coding tool such as Claude Code or Codex.
Prerequisites: Experienced software developers. Comfortable debugging, reviewing pull requests, and working in a team.
Tools and accounts: The course uses Claude Code or Codex as the main coding tool, with a second tool for comparison. The course setup provides six months of access to the selected 5x plan, using either Claude Code or Codex depending on final setup.
Fees (excl. GST): S$1,710 per participant. Minimum 10 participants to start.
All quoted prices are in SGD and are before 9% GST. Invoicing terms available.