Work with me
Facial systems & performance services
Training and consulting for teams in film, games, AI, and robotics. Team education, pipeline alignment, data infrastructure, and model evaluation.
Teams I’ve worked with
Who I work with
Teams building, fixing, or evaluating faces, on screen, in models, and in robots.
Film & games
For animation, rigging, and character teams.
- Rig and blendshape reviews
- Lipsync and speech systems
- Expression systems for creatures and nonhuman characters
- Team training in FACS and facial anatomy
AI & ML
For teams building face-tracking, avatar, and generative video models.
- Expression taxonomies and annotation standards
- Annotator training
- Data collection and capture planning
- Expert review of model output: expression and lipsync
Robotics
For teams designing faces for humanoid and stylized robots.
- Expression and pose libraries for real-time control
- Shape audits and co-activation rules
- Lipsync and speech shapes
- Workshops on facial anatomy and FACS
Beratung
For teams designing, fixing, or evaluating facial systems.
Expression data
Entwurf und Bewertung von Datensystemen für den Gesichtsausdruck.
- Strategie der Datenerhebung
- Beschriftungslogik
- ROM-Design für zuverlässiges Training und Analyse
Rig architecture
Entwurf und Korrektur von Gesichtssystemen für ausdrucksstarke Zuverlässigkeit.
- Bewertung von Rigs, Blendshapes und Korrekturmitteln
- Leitlinien für Formgranularität, Splits und Erweiterungen über das Standard-FACS hinaus
Speech dynamics
Bewertung und Entwurf von sprachgesteuerten Gesichtsbewegungssystemen.
- In-house phoneme and viseme strategies
- System-Audits
- Schulungen für automatisierte und hybride Pipelines
Training & workshops
Spezialisierte Ausbildung für Teams, die Gesichtssysteme bauen.
- FACS, facial anatomy, and lipsync instruction for animation and tech
- A shared language for evaluating rigs, shapes, and facial motion
- Formats from focused lectures to multi-day workshops
Selected work
Film & TV
Hier Provided likeness and de-aging direction and curated training data for the AI de-aging of Tom Hanks, Robin Wright, and Paul Bettany.
Beobachte den Himmel Designed the evaluation framework and quality metrics for expression and lipsync in AI visual dubbing.
Jane (Apple TV+) Developed the expression repertoire for Graybeard, the series’ CG chimpanzee: more than 300 emotional reactions. Covered in The New York Times.
Gigi & Nate Developed a FACS-based expression system for Axis Studios’ digital double capuchin. Covered by Animation UK.
AI & face tracking
Meta, Quest Pro face tracking Founding member of the face-tracking team. Defined the expression taxonomy and annotation standards for the eyes, mouth, and full face, and trained the outsourced annotation teams that applied them at scale. Shaped the data collection strategy.
Patents Co-inventor on drei Patente für Avatar-Animation and expression tracking.
Robotics
Stylized robot face Co-developed a modular expression and pose library with the art team, built for runtime control by the robot’s AI, which selects and combines expressions in real time to match speech and emotional context.
Mechanical robot face Audited the facial shape set. Defined co-activation rules, meaning which shapes may fire together and which combinations to encourage or restrict, and the shape labels used for speech.
Theme park robotics company Led a three-day workshop on facial anatomy and FACS for a multidisciplinary team.
Training
Studio training Delivered facial anatomy, FACS, and lipsync training for teams at Epic Games, ILM, Netflix, Blizzard, Larian, and Rodeo FX.
Who you’d work with
Melinda Ozel
I’m a facial systems expert working across animation, AI, and robotics. I was a founding member of the Meta Quest Pro face-tracking team and am a co-inventor on three patents in avatar animation. My film and TV work includes AI de-aging on Hier, KI-Lippensynchronisation ein Beobachte den Himmel, and a 300-reaction expression repertoire for the CG chimpanzee in Jane.
Fragen
What problems do teams usually bring you?
Expressions that look off without an obvious cause, lipsync that drifts from the speech, labels that disagree from one annotator to the next, and robot faces that fall into the uncanny valley.
When should we bring you in?
As early as possible. Decisions like your expression taxonomy, shape set, or capture plan are easier and cheaper to get right before data is collected or rigs are built. I also work on fixing systems already in production.
How does an engagement start?
With a short call to understand the project. I follow up with a written proposal covering scope, deliverables, timeline, and price.
What do you need from us to get started?
Examples of what you’re working with, such as rig shapes, model output, data samples, or robot footage, plus your goals and constraints.
How is work priced?
Most projects are fixed scope, with set deliverables and a timeline. Ongoing or iterative work may be billed hourly.
What do deliverables look like?
Depending on the project: annotated review reports, shape and pose specifications, annotation guidelines, evaluation rubrics, or training materials and recordings.
Already have a project in mind?
Or just a specific question. Either way, start here.
facetheFACS@melindaozel.com