Film & Episodic VFX · Applied AI

I turn production bottlenecks into AI tools.

I'm Chris — a CG Supervisor working on film and episodic projects for some of the world's biggest studios. On the floor, I see exactly where pipelines stall. Off the floor, I concept, design and ship the AI tooling that unblocks them — wired into the systems crews already use, not demos looking for a problem.

01

Selected systems

Each of these is a complete pipeline, not a wrapper around a single model — multiple AI systems chained together, each doing the one job it's best at, connected to real data.

N/01 · PRODUCTION TOOL Live

BridgeVFX

On every show, artists wait on supervisors to translate client notes into direction — and supervisors are the busiest people in the building. Bridge reads every unactioned note in Flow, works out what the client actually wants, and paints the answer onto the real frame. Direction gets confirmed in seconds, not days.

IN Flow / ShotGrid client notes OUT visual brief + dept. time estimate
Claude Grounding DINO SAM2 FLUX.1 Fill ControlNet Flow API
Open node →
N/02 · STORY / PRE-VIZ Pre-launch

StoryViz

Storyboarding that grows with the production. Draw a rough sketch on a layered canvas, then walk it up a fidelity ladder — sketch → illustrator → colour → cartoon → photoreal — while the whole board stays in one consistent style, with named characters kept consistent across every cell.

IN rough sketch + shot description OUT consistent boards, at any fidelity
i2i diffusion Layered canvas engine fal.ai Next.js Stripe
Open node →
N/03 · CONSUMER APP Launching soon

NovelViz

An AI reading companion that only ever knows what you've read. Ask questions, explore characters, generate imagery straight from the book in your hands — with spoilers made impossible by design, because the model is never shown a page past your bookmark.

IN the book + your reading position OUT spoiler-safe answers & imagery
Claude Context gating Image gen Next.js Postgres
Open node →
N/04 · PERSONAL AGENT R&D

Jarvis — Voice Second Brain

Talk to your notes out loud. Ask a question and the answer is spoken back in a cloned voice while a 3D knowledge galaxy flies to the source note and lights it up. An agent layer drafts emails, checks calendars and actions tasks — with a confirm step before anything leaves the building.

IN voice + markdown knowledge vault OUT spoken answers, notes, agent actions
Claude Agentic tool use ElevenLabs Web Speech 3D graph engine
Open node →
02

How I work

The hard part of AI tooling isn't the model — it's knowing which problems are worth solving and where a tool has to sit inside an existing workflow to actually get used.

01

Find the real bottleneck

Every tool starts on the production floor — a genuine constraint costing real hours, spotted while doing the job. Never a tech demo in search of a problem.

02

Map the workflow

Break the process into steps. Decide which need human judgement and which don't — then pick the right model for each link in the chain.

03

Build & wire it in

Ship a working tool connected to the systems crews already live in — Flow, review, storage — so it removes a step instead of adding one.

04

Keep humans on approval

AI proposes; supervisors approve. Every pipeline I build has a human sign-off exactly where creative accountability lives — that's why they get adopted.

03

The toolbox

Chosen per job, not per hype cycle — the value is in knowing which tool to reach for, and how to chain them.

Models & AI

Claude API OpenAI FLUX.1 / diffusion Grounding DINO SAM2 ControlNet fal.ai ElevenLabs Agentic tool use

Production & pipeline

Flow / ShotGrid API VFX review workflow Client-note triage Bidding & scheduling data Frame / plate IO Show style pipelines

Product & web

Next.js / React TypeScript Postgres (Neon) Cloudinary Stripe Clerk auth Vercel
Contact

Building the same thing for your pipeline?

Whether it's a production bottleneck that needs a tool, or a team that needs someone who speaks both VFX and AI fluently — I'm easy to find.