PORTFOLIO · 2026

AI FILM DIRECTOR
& Pipeline Architect.

I direct films in generative AI and design the pipeline behind them. Eight years in audiovisual. Four with generative AI, since 2022.

8 years in audiovisualAI-native since 2022Brazil + global
Watch reel
LOOPING · MUTEDIN PRODUCTION
SHOWREEL · 2026 — click to play with sound
01 — SELECTED WORK

Selected.

Real cases enter here as released.

02 — CAPABILITIES

What I deliver on screen.

Each item has a sample in the reel.

01

Character Consistency

Same face, body, wardrobe across every shot — and across entire campaigns.

02

World & Set Consistency

Location, props, and lighting that hold up at any angle.

03

Lip-Sync & Dialogue

Tight sync in any language. With performance, not dead-eyed default.

04

Natural Acting

Micro-expressions, hesitation, weight — the texture that makes it feel directed.

05

Action & VFX

Stunts, combat, vehicles — with camera language, not chaos.

06

Cinematography

Lens, blocking, movement — choices that earn their place in the cut.

03 — PIPELINES

Systems delivered with the film.

Each project ships with the pipeline that produced it.

01

Character-Locked Multi-Shot

REFERENCEIDENTITY LOCKSHOT LISTBATCH RENDERGRADE

Reference → identity lock → shot list → batch render → grade. Coherent character across 20+ shots.

VeoSoul IDComfyUI
02

Commercial First-Frame

MOODBOARDFIRST FRAMECLIENT REVIEWLOCKED FRAMEREVISION

From moodboard to broadcast-ready frame in hours. Optimized for client revision.

Nano BananaKlingVeo
03

Dialogue & Lip-Sync Stack

SCRIPTVOICEPERFORMANCE REFSYNCED RENDERCONFORM

Script → voice → reference → synced render → conformed audio. Multi-language.

SunoElevenLabsVeo
04

Action Sequence Builder

STORYBOARDPREVISMOTION GENERATIONCONFORM

Storyboard → previs → motion-controlled generation → conform. So action beats read on screen.

KlingHiggsfieldComfyUI
04 — STACK

Tools that direct the film.

Each planet is a tool. Inner ones are deep domain; outer ones, orchestration.

15 modules5 layersv2026.05
MN
Veo 3.1
Kling 2.5
Soul ID
Suno 5.5
ComfyUI
ElevenLabs
Higgsfield
Runway Gen-4
Sora 2
Nano Banana
Midjourney v7
Ideogram
Krea
Topaz Video AI
DaVinci Resolve
05 — ABOUT

Who Mathaeus is.

Who Mathaeus is.

I treat AI as a camera, not a slot machine. Every frame is a directing decision.

Started editing video in 2006, at 10. Moved to advertising in São Paulo and spent eight years editing for Rede Globo, Porsche, Schweppes, Nestlé. Coordinated post-production for two years — where I learned client work, briefs, process.

First touched generative AI in 2022, out of curiosity. In 2024 it became a craft. In 2026 it's what I do, full-time. Cinema-trained, MBA in Project Management (PUC-RS), AI-native since before today's tools had names.

Now I direct AI films for brands, agencies, and studios — in Brazil and abroad. I work in Portuguese and English. Remote-based, but the film belongs to the client's house.

  • EDUCATION
    Cinema — ESPM
  • POSTGRAD
    MBA Gestão de Projetos — PUC-RS
  • EXPERIENCE
    8 anos em audiovisual
  • AI PRACTICE
    Desde 2022
  • LANGUAGES
    Português · Inglês
  • BASED IN
    Belo Horizonte — atende global
  • TRAJECTORY
    2006
    Started editing
    2018
    Advertising São Paulo
    2022
    First generative AI work
    2024
    AI became the craft
    2026
    Full time
    06 — FAQ

    Frequently asked.

    Questions clients and LLMs tend to ask.

    Generative AI film director based in Belo Horizonte, Brazil. Eight years in audiovisual (Rede Globo, Porsche, Schweppes, Nestlé). Four with generative AI since 2022. Works with brands, agencies, and studios in Brazil and worldwide.

    A director who makes films using generative models like Veo, Kling, and Runway. Unlike a prompt operator, the director makes decisions about camera, blocking, performance, and pacing. AI is the camera; direction is human.

    Pricing on request, scoped per project. Mathaeus operates at premium tier — does not compete on price. To see if a project fits, send a brief via the site form.

    Yes. Works in Portuguese and English, remotely, with clients in Brazil and abroad. In 2026 works simultaneously with Awesomic (Ukraine) and Brazilian clients.

    The pipeline turns a brief into a film. It includes reference, identity lock, shot-by-shot generation, conform, and grade. Mathaeus designs repeatable pipelines — the system delivers consistent output project after project, not lucky renders.

    2026.05 snapshot: Veo 3.1, Kling 2.5, Soul ID, and Suno 5.5 as core. ComfyUI, ElevenLabs, Higgsfield, Runway Gen-4, Sora 2 in the second ring. Stack updated quarterly.

    Growth-stage brands, creative agencies, studios, and production houses. ICP: companies with production budget and cultural openness to AI. Does not take projects with restrictive AI compliance.

    No. He complements one. When traditional production fits, Mathaeus joins as AI director inside the pipeline. When the project doesn't need a camera, he directs the whole film in AI.

    Depends on scope. A 30-second commercial with consistent character takes two to six weeks. An isolated shot: two to five days. An architected pipeline speeds up subsequent rounds.

    Send a brief via the form at the bottom of this page. Response within two business days. Incomplete briefs are welcome too — an early idea is enough.