ESPN: Women’s March Madness

Athletes we couldn't shoot — rebuilt from found footage with AI and ML, dropped into CG worlds.

Case Study

VFX Supervisor ‍ - Team of 5 - Broadcast Promo - Mirada 2026

THE CHALLENGE:

The promos had to feature specific athletes, but we had no access to any of them — no shoot, no capture, nothing. Everything had to be built from found footage of the athletes, which then had to live convincingly inside moody, dramatically-lit CG environments it was never shot for.

MY ROLE:

I developed the core technique and supervised a team of 5, including supervising the VFX on a live shoot used to capture some of our elements and extras.

WHAT I BUILT:

I sourced footage of each athlete delivering the right performance, isolated them, and used AI and machine-learning tools to relight and integrate them into our CG worlds — with a hybrid AI approach driving some of the environments as well. Every shot in the campaign is a blend of found footage, CGI, and generative AI.

THE RESULT:

A polished promo campaign featuring athletes we never had access to — cut together as convincingly as a purpose-built shoot, and an early, production-proven use of AI and ML to solve a problem traditional VFX couldn't.

ESPN: Women’s March Madness Anthem

ESPN: Women’s March Madness Shot Builds