Filmmaking, AI video, post workflows

I help creative teams adopt AI workflows, from demo to deployment.

This portfolio is built to answer a customer-facing question: how do you take generative video, spatial review, creative QA, and rollout decisions and turn them into a workflow a creative team can actually trust, demo, and use?

Read the portfolio Film proof 90-sec walkthrough Team workshop Evidence Demo script Outreach copy

Controllable video S4 documentary DaVinci Resolve Spatial review Creative QA Enterprise rollout

The creative proof comes first.

The technical repos only matter if the creative anchor lands first: co-directing S4, working in DaVinci Resolve, and using AI-assisted post decisions in service of story, continuity, and delivery. The dedicated film proof page makes that production context visible before anyone gets to repo depth.

Film

S4

Feature documentary experience with end-to-end creative judgment, collaboration, and finishing pressure.

Audience

2M+ Views

A JRE clip from the project reached a broad audience, grounding the work in real media impact.

Post

Resolve Workflow

Daily DaVinci Resolve practice connects AI outputs to editing, review, color, and delivery realities.

AI for Story

ML-Assisted Film

De-aging, green screen cleanup, and neural rendering are framed as storytelling tools, not standalone demos.

Open film proof Open walkthrough See the workflow portfolio See validation

Three forwardable surfaces.

This is the recruiter and hiring-manager path: one page for creative proof, one page for the demo moment, and one page for how the workflow would be taught to a real team.

Film proof S4, Resolve, and story-first AI decisions framed as production reality. Walkthrough The recording-ready page for a 60 to 90 second customer-facing demo. Workshop A training surface that shows how the workflow becomes team behavior. Demo deck A compact pre-sales layer a recruiter or account lead could forward.

Proof now.

As of April 11, 2026, the featured repos below are green in GitHub Actions. That means the visible technical trust gap has shifted from red badges to public creative media.

CI Green cinematic-controlnet Prototype path now passes on GitHub runners

Lint, typecheck, docs, and the multiframe regression test are green on main.

CI Green splat-web-viewer Node 24 toolchain plus live Pages demo

Typecheck, lint, test, build, and Pages now present as a clean browser demo path.

CI Green video-gen-eval 28 passed, 1 skipped locally plus green CI

The adapter handoff path, score aggregation, and mypy gate are all clean.

CI Green cloud-inference-platform Router logic plus kernel path backed by green CI

Tests, docs, lint, and typecheck now pass without the old dependency and typing traps.

Four customer-facing workflows.

Each case study starts from a creative or enterprise problem, then maps back to technical exploration in the supporting repos.

01

Cinematic Control

Give directors repeatable control over camera, motion, force, style, and continuity.

02

3D Scene Review

Turn spatial assets into shared references before animation, previs, or generated video.

03

Creative QA

Explain why generated clips pass or fail review with scorecards and plain-language rationale.

04

Enterprise Rollout

Make creative AI workflows reliable enough for latency, cost, observability, and support needs.

Built to be demoed.

Concept diagrams give a fast walkthrough of the workflow logic, and the placeholder media below reserve polished public surfaces without exposing raw production paths or pretending the approved visuals already exist.

Still slot S4 or filmmaker-first frame

The first public proof image for the film anchor, outreach screenshots, and film-proof story.

Timeline slot Resolve editorial reality

The clearest way to prove real post-production context without needing to publish motion first.

GIF slot Workflow loop

A fast recruiter-scan visual for GitHub, the landing page, and any short application surface.

Reel slot 60 to 90 second demo

The main public asset for portfolio sharing once one approved edit is ready to publish.

Technical exploration underneath.

The showcase leads with workflow value, then lets the underlying repos show technical exploration across video, 3D, evaluation, and infrastructure.

cinematic-controlnet Controllable cinematic generation and shot workflow architecture. splat-web-viewer Browser-based Gaussian splat review and spatial media tooling. video-gen-eval Video generation evaluation reframed as creative QA. cloud-inference-platform Deployment reliability, routing, latency, and observability.