Intro to Agentic Engineering
Overview
A taster for our ML engineer apprenticeship. Over one evening you will work with a coding agent on a small project, build a clearer mental model of what is going on when an agent writes code, and use that model to think about effective harness engineering.
Background
For most of computing's history, programming has been an act of translation. You understand the problem in human terms, design a solution in abstract terms, then render it in a syntax the machine can execute, and each step adds friction. That friction is collapsing. Developers increasingly say what they want built rather than how to build it, and the machine handles the implementation while the human supplies intent, architecture and judgment.
The change has come in stages, from autocomplete to inline suggestions to chat interfaces to agents that can plan a multi-file change, run the tests and open a pull request without anyone typing a line of code. By early 2026 most professional developers were using coding agents regularly and a large share of new code was being written by them. This session is about what that means for the way you work.
Who is this for
Experienced software engineers who have used a coding agent, whether in the terminal, such as Claude Code or Codex, or in an editor such as Cursor or VS Code.
What we will cover
1. Working with a coding agent on a real task
2. What changes about the engineering process, and what doesn't
3. ML engineer apprenticeship Q&A
What to bring
A laptop. If you already use Claude Code, Codex or Cursor, bring it. Open-source coding harnesses such as OpenCode work too. If you don't have a subscription, we can give you access to a model API that you can plug into whichever harness you're most comfortable with.