Beyond Vibing Academy · by Groxio

The whole path.
With your team.

Intelligence with intention. Agentic coding, with guidance.

The same four modules, as a private cohort for your Elixir team. Start a conversation with Bruce and Paulo about your codebase, the people learning and the work ahead.

We start with a short call, then a proposal.

Teams we have worked with InervaHiiveErlang Solutions

Developers pairing at a laptop
One course · four modules

From the first story
to reviewed automation.

Four modules, sold together and taught live. The outcomes guide the teaching; the order of lessons can flex.

01 · Coding

Specify the work

Turn a request into small stories and the context an agent needs. When an attempt fails, explain why.

You can show: a clear story, its context and an explanation of a failed attempt.

02 · Steered coding

Judge what comes back

Read agent output against design principles. Discard what does not fit, and explain the remaining system at a whiteboard.

You can explain: the design you kept, the output you rejected and why.

03 · Verified coding

Prove it works

Put verification gates in order of cost. Recognize common AI-generated failures and catch them before review.

You can show: the checks that support a change and the failures they catch.

04 · Automated coding

Automate what has earned it

Run selected automations with rules for adding or reducing them. People still review changes and merge pull requests.

You can explain: what runs automatically, what needs approval and when to pull back.

Scout → Plan → Build → Verify → Review

The same loop throughout. People approve the plan and review the change, including automated work.

The whiteboard test

Can you explain it
without the agent?

How does the system work? Why is it shaped that way? Where would it fail? The point is to explain the decisions behind the code you ship.

You build with guidance, and keep responsibility for what you build.

Who it is for

You already code.
Now direct the work.

For Elixir teams that want a shared way to specify, judge and verify agent-assisted work.

Shared context

Give agents the product knowledge they need to do useful work.

Shared judgment

Agree on what to keep, what to reject and why.

Shared verification

Make the evidence behind a change clear to the people reviewing it.

This course assumes you already write software. It does not teach programming from scratch.

Your teachers

The engineering.
The context.

Bruce Tate
Engineering with agents

Bruce Tate

Founder of Groxio. Leads the technical teaching: directing, judging and verifying AI-assisted work.

Paulo Valim
Context and stories

Paulo Valim

Groxio partner. Teaches context, stories and verification, connecting the work to the business it serves.

A private cohort

The Academy,
for your team.

Talk with us about the people learning, your codebase and the outcomes you need. We start with a short call, then a proposal covering scope, schedule and price.

The same four modules

  • Specify the work
  • Judge what comes back
  • Prove it works
  • Automate what has earned it
  • Price by proposal
Questions, answered

Before you join.

Is this the public cohort?

This is a private cohort for your team. We start with a short call, then a proposal.

Can we work on our own codebase?

Tell us about your system and learning goals in the inquiry. We agree on suitable work together before scheduling.

Who should contact you?

The person coordinating the team's learning. Include the company, team size and what you want the team to learn.

Keep the speed.
Build the judgment.

A shared way to build with agents, with people responsible for the result.