@mattpocockuk: My last AI Coding Cohort was the most successful course I've ever released. 2,500+ students worked with Claude Code for…

X AI KOLs Following Products

Summary

Matt Pocock announces version 2 of his AI Coding Cohort, a two-week course teaching engineers how to effectively use coding agents like Claude Code on real projects, starting June 1st.

My last AI Coding Cohort was the most successful course I've ever released. 2,500+ students worked with Claude Code for 2 weeks, building a real app with AFK agents and software fundamentals. So, we made version 2: - Use any coding agent you like - Updates for every skill - Uses Sandcastle for AFK agents Feeling behind the curve? How would you like to be... ahead of it? Starts June 1st. Get your seat now:
Original Article
View Cached Full Text

Cached at: 05/19/26, 02:41 AM

My last AI Coding Cohort was the most successful course I’ve ever released.

2,500+ students worked with Claude Code for 2 weeks, building a real app with AFK agents and software fundamentals.

So, we made version 2:

  • Use any coding agent you like
  • Updates for every skill
  • Uses Sandcastle for AFK agents

Feeling behind the curve? How would you like to be… ahead of it?

Starts June 1st. Get your seat now:


AI Coding for Real Engineers

Source: https://www.aihero.dev/cohorts/ai-coding-for-real-engineers-m0k0w They open a chat box, paste a vague task, and hope the model figures it out.

That works for toy changes. It falls apart when the codebase is real, the work spans files, and the result needs to survive review.

AI coding is not just prompting. It is engineering with a new teammate that needs context, constraints, checks, and a workflow.

Before you ask an agent to change code, you need to understand the system around the change.

You need to know what matters, where the seams are, what can break, and how you will know the work is done.

That means using AI to explore the codebase, ask better questions, map the relevant files, and turn a fuzzy request into a task the agent can actually execute.

The best AI coding work is not a giant one-shot prompt.

It is a loop:

  • give the agent a clear task
  • constrain the scope
  • let it work
  • inspect what changed
  • run checks
  • feed the result back in
  • repeat until the work is real

This course teaches that loop directly. You will learn how to make agents useful without handing them the keys to wreck your codebase.

AI Coding for Real Engineersis a two-week cohort for developers who want to use AI coding tools on real production-style work.

We will use Claude Code heavily, but the point is bigger than one tool. You will learn workflows that transfer across coding agents: context gathering, planning, steering, feedback loops, autonomous runs, and human-in-the-loop review.

You will get a structured set of lessons, exercises, a real project repo, live office hours, and recordings you can keep.

Each week unlocks a set of practical modules. You can follow along live, work async, or catch up when your schedule allows.

Full Schedule: What you’ll be learning, doing, and building each week

Each Monday, you will get access to that week’s lessons. You can rip through the material, revisit lessons as needed, and bring questions to office hours.

Pre-course: Getting to Know Claude Code

Before the cohort starts, you will get pre-course lessons that help you get comfortable with Claude Code and the course workflow.

If you are brand new to Claude Code, this gets you oriented. If you already use it, this gives you a clean baseline for the exercises.

Week 1: Fundamentals, Steering, and Planning

Week 1 is about making the agent understand the work before it touches the code.

You will learn how to:

  • explore a codebase with AI
  • find the files that matter
  • write better task briefs and PRDs
  • use steering docs and project context
  • turn vague ideas into executable plans
  • work in small loops instead of giant prompts

Week 2: Feedback Loops, AFK Agents, and Human-in-the-Loop Patterns

Week 2 is about getting useful work out of coding agents while keeping engineering judgment in the loop.

You will learn how to:

  • build feedback loops with tests, types, and runtime checks
  • run agents against real tasks
  • review and recover from bad agent work
  • decide what can run AFK and what needs supervision
  • use human-in-the-loop workflows for quality and taste

Your future with AI coding is inside

The goal is not to make you dependent on a specific tool.

The goal is to help you build judgment and workflow.

You will leave with a practical way to use agents on real code: planning, execution, validation, review, and iteration.

Learn on your schedule, from anywhere

The cohort is designed for async participation.

Office hours are live on YouTube, recorded, and available afterward. If the live times do not work for you, you can still follow the course and keep the recordings.

  • Two weeks of structured cohort lessons
  • A real project repo for the exercises
  • Pre-course setup lessons
  • Live YouTube office hours
  • Office hours recordings and transcripts
  • Discord access for cohort discussion
  • Lifetime access to the material from this cohort run

Which Claude plan do I need?

You should have access to Claude Code. The course uses Claude Code heavily, and the pre-course material helps you get set up.

Can I use a different AI coding tool?

Yes, but Claude Code is the main tool used in the lessons. The workflows transfer, but if you use another tool you may need to translate some steps yourself.

Do I keep access forever?

Yes. Once you purchase the cohort, you keep access to the lessons, exercises, recordings, transcripts, and office hours recordings for this cohort run.

What if I can’t attend live or have timezone conflicts?

That is fine. Office hours are recorded, and the course is designed so you can work through it async.

What’s your refund policy?

If you buy the cohort and it is not a fit, contact support and we will help.

More questions?

Email support and we will point you in the right direction.

Similar Articles