AI Creative Director

Don Allen Stevenson III Portfolio

donalleniii.me documents a path from DreamWorks Animation to AI Research Lead at the AI film studio Asteria and MasterClass instructor, including an OpenAI Sora-made nature documentary and a six-month Meta Connect collaboration that ended in a joint keynote with Mark Zuckerberg.

Generative AIAI Animation PipelinesCreative TechnologySpatial Computing

Location

Remote-first, in the most literal sense — no idea where.

Skills

Generative AIAI Animation PipelinesCreative TechnologySpatial Computing

What makes it work

A breakdown of the choices that make this portfolio stand out.

1

Traditional animation experience as the foundation for AI film work

Coming from DreamWorks Animation gives weight to an AI research lead role at an AI film studio: he knows the production workflow the new tools have to fit into.

2

Named, verifiable collaborations rather than vague brand mentions

A documentary made with early Sora access and featured on OpenAI's own platform, and a six-month Meta engagement culminating in an on-stage keynote, are specific, checkable claims — not "worked with leading AI companies" language.

3

Shipping small functional tools, not just giving talks

Alongside keynotes, the site lists concrete built tools (a real-time ASCII art engine, a browser-based Gaussian-splat converter) — evidence of ongoing hands-on building.

What AI Creative Directors can take from this

Specific, actionable tips to apply to your own portfolio — no generic advice.

  • If you're pivoting into an AI-creative title, show the traditional production-pipeline experience that makes the pivot credible.

  • Name the specific collaboration and its outcome instead of a generic list of company logos.

  • Keep shipping small, concrete tools even as your public profile grows — it keeps the portfolio grounded in building, not just appearing.

Projects

Don Allen Stevenson III's site shows him on stage and at the build bench. There's a Sora film OpenAI featured, a live Meta Connect demo of his AI clone, and small tools he ships and open-sources himself.

Creative Demoadvanced2024 (early Sora access)

Beyond Our Reality

OpenAI SoraGenerative videoCreature design

An AI-generated nature documentary about impossible hybrid creatures, made with early access to OpenAI's Sora. His site calls it one of the first films created with Sora. OpenAI's 'Sora: first impressions' post quotes him using the model to prototype hybrid creatures before he builds them out as 3D characters.

Early-access work that the model's maker chose to showcase. It shows he turns a new tool into a finished creative piece, not just test clips.

  • One of the artists in OpenAI's Sora first-impressions post

  • Uses Sora to prototype creatures before building them in 3D

  • Listed on his site as one of the first films made with Sora

Creative DemoadvancedSix-month collaboration; keynote September 2024

Meta Connect: Creator AI demo with Mark Zuckerberg

Meta AI StudioCreator AILive keynote demo

His site says he worked alongside Meta's teams for six months on the behavioral models behind AI Studio, the platform creators use to build AI versions of themselves. At Meta Connect, Zuckerberg held a live video conversation with Stevenson's AI clone while Stevenson stood on stage.

He helped shape a consumer AI product from the creator's side, then demonstrated it live at the company's biggest keynote. That is product collaboration, not just endorsement.

  • Six months with Meta's teams on AI Studio behavioral models (per his site)

  • Live on-stage AI video call with Zuckerberg at Meta Connect

  • Covered by TechCrunch

Open SourceintermediateOpen source on GitHub

SharpSplat

PythonPyTorch (MPS)Apple ml-sharpGaussian splattingmacOS

A small Mac app that turns a single photo into a 3D Gaussian splat (.ply), locally on Apple Silicon with no cloud round-trip. It wraps Apple's ml-sharp model and then post-processes the output. It strips camera data that some viewers and After Effects plugins reject, and renders a 24-frame rotating preview.

It packages a research model into a drag-and-drop tool that plugs into real creative software. That is the pipeline work an AI creative lead does for a team.

  • Single image to ~1.18M Gaussians in one forward pass (per README)

  • Output cleaned to the standard 14-field splat format for After Effects, SuperSplat, Luma and Polycam

  • One-line install script sets up Python, the model and the app

ToolbeginnerLive browser tool

GLYPH

Browser (GitHub Pages)Real-time renderingCamera input

A real-time ASCII art engine in the browser. It turns camera feed, imported images and 3D shapes into moving text art, with import, capture and undo controls.

It's a small, finished interactive tool he shipped himself. It backs up the site's claim that he builds working systems, not just demos for talks.

  • Runs entirely in the browser

  • Takes camera, image or 3D-shape input

  • One of several shipped tools on his 'Systems I Build' list

Ready to build your portfolio?

Follow the AI Creative Director roadmap — skills, projects, and timeline to get hired.

AI Creative Director Roadmap