Workshop
One team. One focused intervention.
For a team that needs shared understanding, practical capability, and something working by the end. Your problem, your people, hands on the machines.
- Single team
- Shared understanding
- Working prototype
AI enablement for teams
Your team doesn’t need another AI presentation. It needs a working session.
What this is
Hands-on AI enablement through focused workshops and build-intensive hackathons. Your problems, your people, working software.
Takes about two minutes
Two ways to work
A workshop, or a hackathon.
Both are hands-on. One goes deep with a single team; the other goes wide across several. Scope and structure are set together before anything is booked.
Hackathon
Multiple teams. Build-intensive.
A build-intensive programme that activates several teams at once, produces working solutions, and makes the next investments visible to the people who fund them.
- Multiple teams
- Working solutions
- Next investments visible
For a dedicated Buildathon-as-a-Service engagement—for a company, community or conference—visit Buildathon.eu.
Explore Buildathon.eu
Who this is for
Three situations that end up in this room.
S.01
A mandate without adoption
Leadership announced the AI strategy. The teams still work the old way. You do not need another all-hands — you need a room where people actually build.
S.02
A program stuck at pilots
Decks, pilots, proofs of concept — none of it reached daily work. The stall breaks when a team builds something real on a problem it owns.
S.03
A team that has to ship
Founder or product lead. Your team must build with AI tools now, on the actual product, under real constraints. Watching tutorials is not a plan.
Topics
Four topics. Either format.
These are the topics a session runs on. Any of them can sit inside a workshop with one team or a hackathon across several.
T.01 — People at work
AI Builder
Turn AI-assisted building into a team skill, not a party trick.
- In the room
- The team builds a working product increment on a real problem from your backlog — prompting, iterating, and shipping under review.
- Leaves the room
- A working build, the prompts and decisions behind it, and a team that has done it once and can do it again.
T.02 — Machines and models
Agents
Move agents from slideware to operations.
- In the room
- Design and run an agent against a task your team owns — scope, tools, guardrails, and failure modes, tested live.
- Leaves the room
- A running agent prototype, its operating boundaries written down, and a list of where it must not act alone.
T.03 — Machines and models
Claude
Make Claude daily infrastructure for your team.
- In the room
- Projects, context, and workflows configured around how your team actually works — then drilled on real tasks until the workflow holds.
- Leaves the room
- Working Claude setups per function, shared conventions for use, and habits that were practiced, not described.
T.04 — Machines and models
Codex
AI-assisted engineering that survives code review.
- In the room
- Your developers pair with Codex on your repository — real tickets, real tests, real review standards.
- Leaves the room
- Reviewed changes, the prompts and review notes behind them, and a workflow ready for engineering-lead review.
Public sessions
See the method in the open.
Two practical sessions on building with AI and turning adoption into something people actually do.
Dima Abramov × Jake Lee · Maven
How to Vibe Code with Lovable and Amplitude
Build and instrument in the same flow: product structure, live implementation and Amplitude from the first visitor.
View the sessionDima Abramov × Elena Avramenko · Maven
How to Drive AI Adoption with a Buildathon
A practical session on format, preparation and what makes a buildathon work for technical and non-technical builders.
View the session
Procedure
How a workshop actually runs.
Before the room
- B.01
Qualification
A short intake. Your situation, your team, your constraints — so the room is built around you, not a template.
- B.02
Problem selection
You bring candidate problems from the real backlog. Together we pick the ones worth the room’s time.
- B.03
Setup
Tools, accounts, and access arranged before anyone walks in. The room starts building, not configuring.
In the room
- I.01
Brief
The objective for the day, stated plainly. What “working” means is agreed before hands touch keyboards.
- I.02
Build
Hands on the machines. The instructor builds alongside the team, not in front of it.
- I.03
Review
The output is checked against the objective. Failures are named, not smoothed over.
After the room
- A.01
Handoff
Artifacts, prompts, and decisions documented, so the work survives the room it was built in.
- A.02
Adoption
A follow-through plan: what the team practices next, and who owns making it a habit.
Evidence · Snapshot 2026
The record so far.
- Events hosted
- 50+
- Luma · lifetime
- Community rating
- 4.6/ 5
- Meetup · 2026
Community
2,000+ people
Across Luma, Meetup, workshops and build sessions.
A rounded snapshot of the studio’s workshops and community.
Participant record
What participants said, verbatim.
Recorded after community workshops.
“I attended the Loveable workshop in November, and it was an absolutely fantastic experience! The host was super clear and helpful, and the content was perfect for beginners. Thanks to the workshop, I was able to build my first project in Loveable. It was truly inspiring! ✨”
“Dimitri was super helpful, and he helped a lot during the process”
“I’ve built my first project in Lovable in less than 10 minutes without having a technical background.”
Questions, answered
Asked before every session.
Leaders responsible for making AI adoption real: Chief AI Officers, Heads of Transformation, founders, and the product and engineering leads who have to turn a mandate into daily practice.
A workshop suits one team that needs shared understanding, practical capability, and something working by the end. A hackathon suits several teams building at once, when the goal is working solutions and a clearer view of where to invest next. If you are unsure, say so in the form and we work it out together.
No fixed level is required. AI Builder works for mixed teams; Codex assumes working engineers. The right bar for your team is set during qualification.
The four session topics: AI-assisted building with tools like Lovable, agent design and operation, Claude, and Codex. The topic is chosen around your situation, not the other way around.
Yes — that is the point. You bring candidate problems from your own backlog, and we select the ones worth the room’s time together during preparation. No canned exercises.
On-site at your office or remote. On-site is recommended for a first room; remote works well for follow-up sessions.
Data handling, account access, confidentiality, and any public use of the work are agreed before the session. Wherever possible we work inside your own accounts and tools. Raise specific requirements during qualification.
Together, before anything is booked. Team size, situation, and topic determine the structure and duration. There is no fixed price list on this page.
You complete the short form — about two minutes. If the fit is right, you hear back with proposed next steps and a preparation call.
Get started
Bring a real problem. We’ll bring the room.
Tell us what your team is up against and whether you have a workshop or a hackathon in mind. Your answers go straight to the studio — no sales sequence in between.
About two minutes