Agentic Engineering with Claude Code

Coding agents are getting strong enough to implement entire applications and refactor unfamiliar codebases. At the same time, if not managed properly, the generated output can be poor, inefficient, or plain wrong. This full-day, hands-on masterclass teaches practical AI-assisted coding with Claude Code, focusing on agentic workflows that help you plan, steer, and verify changes efficiently. The day alternates between concise instruction blocks, guided exercises, and short Q\&A sessions.

By the end of the masterclass, you will be able to:

• Turn coding tasks into clear, actionable instructions (goals, constraints, definition of done)
• Curate and maintain usable context for Claude Code while avoiding overload
• Apply simple customization approaches to align Claude Code with personal or project conventions
• Recognize key risk areas (secrets/sensitive data, licensing/IP) and mitigate them in practice
• Use Claude Code’s tools and extensions, and choose suitable mechanisms for a task
• Apply practical agentic workflows for both greenfield and brownfield work, including planning, checkpoints, verification, and session handoff

Who should attend?

Attend if you write Python and want a reliable, repeatable way to work with coding agents, whether you are prototyping new code or maintaining an existing codebase. If you are new to AI-assisted coding, you will learn a clear workflow and guardrails that prevent common failure modes. If you already use Claude Code, we are sure that you will still see something new.

Requirements

• Claude Code installed and an active subscription that enables Claude Code usage (follow instructions here)
• uv installed (follow instructions here)
• Internet connection during the tutorial will be provided

Format

This tutorial is interactive and exercise-driven. Each topic block roughly follows the same structure:

1. Short instruction and live demonstration
2. Guided hands-on exercise
3. Brief discussion and Q\&A