AI training that doesn't suck.

Six sessions on the real stack — not last year's demos. 50% learning, 50% laughing, 100% hands-on. By the end, your team can ship a model, render a video, and — the part nobody teaches — know when not to use AI at all.

FORMAT
In-person + live remote
SESSIONS
6 modules
DURATION
12 weeks
COHORT
4–12 people

/ BUILT FOR

Working professionals — directors, designers, editors, producers — who already ship creative work and want to integrate generative AI into a real, deadline-bound pipeline.

You don't need to be technical. You need to be curious, somewhat skeptical, and willing to make the bad first draft.

/ WHAT YOU LEAVE WITH

  • A working AI image pipeline integrated into your existing stack
  • One custom-trained LoRA in your studio's style
  • Practical literacy on video models — Runway, Veo, Kling, Sora
  • An honest map of what to use, what to skip, and why
  • A finished piece made with everything above

Six sessions, in order.

Each session builds on the last. By Session 4 the team is making genuinely useful images. By Session 6 they've shipped something they're proud of.
01 / 06
GROUNDING — Maps, models, and the mythology to unlearn
Where we actually are with this technology — and where the marketing departments would like you to think we are. The landscape of models, the economics of running them, and the studio-level decisions you'll be making in the next six weeks.
FOUNDATION
02 / 06
IMAGE — From Photoshop muscle memory to prompts that work
Hands-on with Midjourney, Flux, Nano Banana Pro, Seedream, Qwen-Edit. Translating your existing skills (composition, lighting, retouching) into a new vocabulary. Where AI saves hours and where it eats your day.
FOUNDATION
03 / 06
VIDEO — Motion, control, and the new grammar
The video model stack — Runway, Sora, Veo, Kling, Higgsfield. How to think about i2v vs t2v, camera control, character consistency, and the surprising places these models break.
WORKSHOP
04 / 06
CUSTOM — LoRAs, fine-tuning, and style ownership
How custom training actually works. We train one LoRA in your studio's style — together. You leave with a working, reusable model that's actually yours.
WORKSHOP
05 / 06
PIPELINE — ComfyUI, automation, scale
The unsexy infrastructure layer. ComfyUI workflows, API integrations, asset management, batch processing. How a tool stops being a toy and starts being a pipeline.
WORKSHOP
06 / 06
PRACTICE — A real project — end to end
No tutorial. No exercise. Your team picks a real piece, scopes it, and ships it using everything above. I'm in the room. We finish it together.
WORKSHOP

Your existing skills, reframed.

Nothing about generative AI is foreign to a senior creative. Every concept maps to a tool you already use. The vocabulary changes; the craft does not.
What you already know
How it shows up
Photoshop selection masks
Photoshop, since 1990
ControlNet
Lightroom presets and looks
Adobe Camera Raw
Custom LoRAs
After Effects keyframes
Motion control
Camera prompts
Premiere proxy workflows
Working with low-res first
Image-to-video iteration
Mood boards & reference
Pre-production research
Style transfer + IP-Adapter

The current stack, honestly assessed.

This list updates every quarter — because the landscape does. You'll leave knowing which to use, which to skip, and which to keep an eye on.

Image / generation

7 platforms — current stack
Midjourney Freepik Krea Nano Banana Pro Seedream Qwen-Edit fal.ai

Video / generation

7 platforms — current stack
Runway Sora Veo Kling Higgsfield Hailuo Pika

How it runs.

01 / DURATION

12 weeks end-to-end

Six sessions, two weeks apart. Enough breathing room to actually practice between sessions; not so much that momentum dies.

02 / FORMAT

In-person + live remote

On-site at your studio when possible. Live remote works fine for distributed teams; recordings stay available.

03 / COHORT SIZE

Small, on purpose

Four to twelve people. Past that, this stops being a workshop and starts being a lecture.

04 / BETWEEN SESSIONS

Office hours + async

Weekly drop-in. Always-on Slack for the cohort. You hit walls; we get around them within hours, not weeks.

05 / DELIVERABLES

A finished piece

Plus a custom LoRA, a documented pipeline, and a team that can actually use them.

06 / SCOPE

Tailored, every time

The curriculum is fixed. The case studies, the projects, and the LoRA we train are yours.

Let's see if this fits your team.

Thirty minutes. Tell me about your studio, your timeline, and what you're actually trying to do. I'll tell you whether this program is the right move — and what you'd do instead if it isn't.

01 / SCOPE CALL

30 minutes

A real conversation. Your goals, your stack, your team. No deck.

02 / PROPOSAL

Within 1 week

A short, specific proposal. Curriculum tailored to your work, fixed price.

03 / KICKOFF

2–4 weeks out

First session locked. Materials sent. Your team starts before week one.