Learn to build AI tools.
Build your first in 30 days.

Learn to plan, build, test and deploy your own AI tools in a live challenge with me, Owain Lewis. Your first project: an assistant that researches a topic, turns meeting notes into next steps or answers questions using your documents.

I’ll build mine alongside you, explain my decisions and help you work through yours. Follow a code or no-code route, with weekly sessions and help when you get stuck.

Your first project: an AI assistant.

Choose one task for your assistant. Here are four ways you could practise connecting AI to information and getting a result you can check. These are project ideas you can adapt.

Research a topic

Collect sources and prepare a brief you can check, using the questions and standards that matter to your work.

Prepare meeting follow-ups

Turn your notes into decisions, next steps and draft messages for you to review.

Find answers in your documents

Ask questions of the material you work with, and check each answer against its source.

Prepare your morning briefing

Bring together your calendar, priorities and updates so you can see what needs your attention.

Watch me build. Get help with yours.

You’ll see how I choose tools, test an idea and fix a failed approach. Then apply that process to your own project, with feedback in live group sessions.

A starting point you can adapt.

Get code and no-code example builds, tutorials, prompts and specifications in the Build Kit. Start with a working example and change it to suit your job.

Weekly Build Sessions.

See my assistant take shape, including the decisions and mistakes. Show your progress and get feedback on your scope, approach and next step.

Build Clinics for blockers.

Bring a wrong answer, a broken integration or a part you cannot get working. Share your screen and work through it with help. Session times will be confirmed before enrolment.

Practise each step, from plan to deployment.

We’ll help you set up before we start. Each week, you’ll get the material for the next step and time to work on your assistant. The aim is a working version in 30 days, plus a process you can use for your next project.

  1. 01

    Design

    Choose one job.

    Decide what your assistant should do, choose your tools and make a plan small enough to finish.

  2. 02

    Build

    Get a first version working.

    Follow the code or no-code example, connect the pieces and bring your first demo to the group.

  3. 03

    Improve

    Try it with your own work.

    Use real data, check its answers and fix the failures. Add human approval before it sends messages or changes anything.

  4. 04

    Ship

    Get it ready to use.

    Work through the deployment, hosting, security and monitoring your assistant needs. Prepare a working version for Demo Day.

Demo Day.

Show your assistant doing its job and explain how you built it. Compare what worked and what failed, and decide what you would improve next.

Build with an experienced engineer.

Meet Owain at Gradientwork

I’m Owain Lewis, an engineer and director in big tech with 20 years of experience building software.

I’ve owned and operated cloud systems used by millions of people. I run AI Engineer and build tools for my own businesses, including the carousel app I use to create content.

In this challenge, you’ll follow my own assistant build as it develops. I’ll explain why I choose one approach over another, review your progress and help you fix the things that stop your project working.

Owain Lewis

Learn by building with us.

For people ready to move from using AI to building with it. Spend 30 days practising on your own project, with examples to adapt and an experienced engineer to help you make decisions.

Join the waitlist to hear the dates, price and session times. You can decide whether to take part once those details are ready.

BuildTogether

First challenge: AI personal assistant

30-day challenge. Dates and pricing to be announced.

  • Code and no-code examples, tutorials and Build Kit
  • Weekly Build Sessions with Owain
  • Build Clinics to work through blockers
  • Demo Day and feedback on what you made
Join the waitlist

Complete the short form to join. No payment today.
Tools and hosting cost extra.

Before you join.

Do I need to know how to code?

No. Choose a no-code route using visual tools, or a code route using a starter repo and AI coding tools such as Claude Code or Codex. Both have example builds and tutorials. We help you choose and set up before the challenge begins.

Who is this for?

People who use AI and want to learn how to build tools for their own life or work. You can be a first-time builder or an experienced engineer. Bring a task you want to tackle and time to build between sessions; you do not need a finished plan.

Is this useful if I am already an engineer?

Bring your own approach, add more demanding integrations and get feedback on your technical decisions. You can move ahead of the weekly material.

Will you build it for me?

You do the building. I help you choose an approach, review your progress and work through blockers in the live group sessions. This does not include done-for-you development or unlimited one-to-one support.

How much time will I need?

Set aside time each week for the live sessions and focused building between them. We will keep the scope realistic for the time you have. A working version in 30 days is the aim; progress depends on your scope and the time you put in.

Are software costs included?

No. Tool subscriptions, model usage, hosting, domains and other third-party services cost extra. We will help you choose a simple setup for your project.

When does it start, and what does it cost?

Dates, pricing and the live session schedule are to be announced. Complete the waitlist form to hear when they are ready. Joining the waitlist is free and does not reserve a paid place.

What happens after Demo Day?

We compare what worked, what failed and what we would change. Use those lessons to keep improving your assistant. Ongoing development and maintenance are your responsibility.