Yoshi Sakaguchi East Tennessee Est. 1998 Edition v — Broadsheet

Pictures that move

Thirty years of illustration, motion, 3D and game art — and two spent teaching the machine to help without taking the hand out of it.

Graphic design, motion graphics, video and generative AI for Jellyvision, Jackbox Games and Tynker. Now Senior Learning Experience Design at Cirrus Aircraft, leading AI research and development. Every picture on this page is mine.

1998
First 3D character, broadcast nationally
9.4M
Views, Mythicraft reel, Tynker
18M
Employees reached at Jellyvision
Six
Aircraft courses in the Cirrus LMS
01

The Machine Room

The current work

Two years of teaching generative tooling to carry an idea from first sketch to finished product without losing the hand that made it. Not prompts typed into somebody else's website: a stack that runs on my own machine, models I have trained myself, and pipelines built to produce the same answer twice.

A model will happily give you something plausible and wrong. Knowing which is which is still the job.

I came to it as an artist, which is the part that matters. Thirty years of drawing means I can tell at a glance when a hand has six fingers, when a light source has wandered between two shots, and when a thing is merely impressive rather than right.

The rest is engineering: reference sheets, trained weights, automated gates, and the patience to run a shot again. It is the same craft it always was. The brush is just faster, and it argues back.

Where I stop

Client material does not go to a third-party service. A real person's likeness does not get generated without their say-so. And nothing ships that I have not looked at frame by frame — a tool that makes a thousand images still needs somebody who can say which one is right.

Specification — what runs here
Machine Apple silicon, local and offline. No upload, no third party, no client data leaving the building.
Front ends ComfyUI, InvokeAI and AUTOMATIC1111 over one shared model library.
Image Flux.1 and Flux.2, SDXL, Qwen-Image, Z-Image, SD 1.5 — chosen per shot, not per habit.
Video Image-to-video and restyle passes, driven shot by shot rather than clip by clip.
Training LoRAs trained on my own reference sheets, so a character stays the same character across hundreds of frames.
Tooling A desktop app I wrote: finds and installs models from Civitai and Hugging Face, queues restyle and video jobs, and keeps the kit portable on an external drive.
In service The Designery, 2024—25: renders turned around inside a meeting, and client walkthroughs in VR. Cirrus Aircraft, 2025—now: R&D inside a regulated training programme.
01 — InterviewWhat the expert actually said, captured once.
02 — DesignOne layout, made by hand, that the machine then follows.
03 — GenerateModels, LoRAs and templates filling it out at volume.
04 — VerifyA gate that runs itself. Nothing ships that cannot be traced.
02

Moving Pictures

Ten films · drag sideways

Directing, shooting, cutting and animating. Nothing is requested from YouTube or Vimeo until you press a play button — the frames below are served from this page.

Mythicraft title card — custom vector characters over a Minecraft capture.
I

Mythicraft

Minecraft capture, custom vector graphics, After Effects. 9.4M views.

Employee Healthier Living — title frame.
II

Employee Healthier Living

Vector assets created, edited and animated in After Effects.

Wellpoint Castlight — title frame.
III

Wellpoint Castlight

Captured footage, edited and animated in After Effects.

Tynker for Fun — title frame.
IV

Tynker for Fun

Captured footage, edited and animated in After Effects.

Agent Recall — title frame.
V

Agent Recall

Gathered and edited footage, crafted with Final Cut Pro.

Gears of Time — title frame.
VI

Gears of Time

Edited gathered footage, refined with Final Cut Pro.

Training Day — title frame.
VII

Training Day

Edited and finished for the Tynker channel.

Coding Cup — title frame.
VIII

Coding Cup

Event reel — footage gathered, cut and titled.

Tynker On The Go — title frame.
IX

Tynker On The Go

Product spot for the mobile app.

Tynker Junior — title frame.
X

Tynker Junior

Spot for the youngest end of the product range.

Film I of X
03

Plates

Forty-two · tap to enlarge
04

Dossier

The full index
05

The Record

All six references

I started in 3D in 1998, modelling characters for films that went out on national television. That taught me to think in space before I think in layers — and it is still how I plan a frame, a lesson or a page.

Since then: nine years at Jellyvision turning benefits enrolment into something watchable, a contract month at Jackbox drawing two hundred props, three years at Tynker running graphics for a product children actually used. Different rooms, same job — make the thing clear, then make it worth looking at.

The newest part of the practice is the generative one, and I treat it the way I treated 3D in 1998: learn the machine properly, then use it to do the work I already knew how to do, faster and at a scale I could not reach alone.

Yoshi Sakaguchi
The author · East Tennessee
The record — 1994 to now
2025 — now Cirrus Aircraft Senior Learning Experience Design. Lead of AI research and development; six aircraft systems courses delivered into the LMS.
2024 — 25 The Designery Designer. Built an AI imaging workflow that turned renders around in minutes, and ran client walkthroughs in VR.
2020 — 23 Tynker Game & marketing graphics lead. Managed international editors and artists; 9.4M views on the Mythicraft reel.
2020 Jackbox Games Game & marketing artist, contract. The Devils and the Details, Talking Points, Party Pack 7.
2011 — 20 Jellyvision Lead production artist. Nine years of benefits-enrolment design reaching 18 million employees and 114 of the Fortune 500.
2009 — 11 Webfoot Technologies Lead game designer. Nintendo DS — You Don't Know Jack, Fancy Nancy, The Biggest Loser.
1999 — 2006 Westcreek Studios Lead 3D modeler. Two nationally broadcast films, 300+ characters, and the modelling division built around them.
1994 — 98 Northern Illinois University BFA, Timed Arts.
“Jellyvision was the very lucky home for Yoshi's artistic excellence for almost nine years. He knows how to handle high volume, high pressure project work as well as iterative product work — two different mindsets.”
Amanda LannertJellyvision — CEO
“Someone who reliably brought clever ideas to the table, and nailed the landing when it came to execution, often on his first stab at it. Can't recommend him enough.”
Mark RaderJellyvision — Marketing
“He was always on the forefront of figuring out how to adapt our art for the constantly-evolving web landscape, and how to streamline the process of creating it.”
Katie LauffenburgerJellyvision — Design
06

Correspondence

Reply within a day

Hiring, commissioning, or just working out whether I'm the right fit — write to me directly and you'll get a person, not a funnel.

yoshi.sakaguchi@gmail.com

StatusOpen to work

BasedEast Tennessee · remote, hybrid or on site

DoingGraphic design · motion · video · 3D · generative AI

Elsewhere LinkedIn · YouTube · Vimeo

© 2026 Yoshi Sakaguchi v1 · v2 · v3 · v4 · v5 Set in Bodoni Moda, Chivo and Space Mono