A graphics engineer who decided clothes were just physics waiting to be solved
Watch a 3D dress fall onto a virtual body inside z-emotion's software and the seams pull, the hem swings, the fabric settles with a little overshoot before it comes to rest. That overshoot is the whole company. It is the difference between a flat picture of a garment and a garment that behaves. Dongsoo Han has spent his career chasing that behavior, first in hair, then in cloth.
Han is the founder and chief executive of z-emotion, a 3D garment simulation company he started in 2017. The pitch is direct: let fashion brands design, fit, render, and animate clothing entirely in 3D, before a single physical sample is cut. The company's tagline, "Forward Fashion Faster," reads like an engineer's complaint about how long the old way takes.
Before any of that, there was a ponytail. Working inside AMD's GPU Tech Initiative Group as a software engineer and researcher, Han built real-time physics for the hard problems of computer graphics - rigid bodies, fabric, hair, glass - all running on the graphics card instead of the processor. The hair work became the core of AMD's TressFX, which shipped in the 2013 reboot of Tomb Raider and was described at the time as the first playable hair simulation in a video game. Lara Croft's hair moved with gravity, wind, and momentum, every strand checked against the others so none passed through her shoulders.
"A thread moving through 3D space behaves a lot like a strand of hair."
That observation is the hinge of his whole career. A head of hair is thousands of thin strands, each one a chain of links responding to force. A woven garment is thousands of threads doing something similar, only knitted into a surface. Han had already written the textbook on the first problem - literally, the chapter "Hair Simulation in TressFX" in the GPU Pro 5 graphics programming series. The second problem was the same math wearing different clothes.
The TressFX work was not a parlor trick. It treated each strand as a chain of links and let gravity, wind, and the movement of the head push them around, with collision detection making sure no strand slipped through another or through a solid surface. The reported scale - on the order of 19,000 strands and roughly 0.22 million vertices, computed in under a millisecond on high-end hardware - is the part that mattered for what came later. Real-time means a designer changes something and sees the result now, not after a render queue. Carry that expectation into fashion and the whole tempo of designing a garment changes.
Han has described the advantage of his background plainly. Most fashion CAD software was built for an earlier era and a narrow professional audience, with conventions baked in over years. Coming at the problem from the gaming and visual-effects side, he was not bound by those conventions. He could ask what a fashion design tool would look like if you built it today, on current hardware, for an audience that now includes game studios and metaverse platforms as much as apparel houses.
From Silicon Valley physics to a Seoul fashion startup
Han's path runs through the University of Pennsylvania, where he earned a master's degree in computer science between 2007 and 2010. His skill list reads like a graphics engineer's resume - CUDA, OpenGL, C++, simulation, computer graphics - the toolkit for making pixels obey the laws of motion. After AMD, he founded ZelusFX and built the Zelus Engine, an avatar system that became the groundwork for the digital bodies z-emotion now drapes clothing onto.
The leap from gaming to garments is less of a leap than it sounds. Both industries care intensely about how surfaces look and move under light. The fashion industry, though, had largely been doing it by hand - sketch, cut, sew, photograph, repeat. Han's bet was that the same simulation pipeline that dressed game characters could compress that loop, and do it with enough fidelity that a brand would trust the screen instead of the sample.
The products: a 3D wardrobe for an industry
z-emotion's tools arrive as a connected set. z-weave is the flagship - a 3D garment design CAD that lets designers build clothing as simulated cloth rather than flat patterns. z-fit is a virtual fitting plugin for trying garments on digital bodies. z-maya hooks the engine into Autodesk Maya, so animators and content creators can run realistic cloth inside a tool they already know. And zeavric, launched in 2023, is an intuitive 3D fashion design studio for characters, garments, accessories, AR environments, and collectibles.
The strategy behind zeavric was to break, in Han's words, "the tradition of what a fashion 3D CAD used to be." Most fashion CAD tools were built years ago for a narrow professional audience. By building on newer technology, z-emotion could aim the same engine at the markets crowding in from the side: gaming, the metaverse, digital humans, and entertainment. A single physics core, many industries.
Big names on the rack
The proof points come with logos. Projects with brands such as Louis Vuitton and Nike show how z-emotion folds technology developed for gaming and film into fashion's actual business processes. The company calls this convergence, and it is the part that makes the engineering matter commercially: a luxury house and a sportswear giant care about the same thing, which is whether a digital garment is convincing enough to design with.
The investor list reflects that crossover appeal. z-emotion has drawn strategic backing from LG Electronics, Naver, HTC, and Shima Seiki - a roster that spans consumer electronics, internet platforms, virtual reality hardware, and industrial knitting machines. The Shima Seiki connection is telling: the company is one of the world's best known makers of computerized knitting machines, the physical end of an industry that z-emotion is digitizing at the design end. In 2022 the company also stepped into digital collectibles, introducing a digital fashion NFT on the CYPHRLY platform.
The technical glue in all of this is connectivity. z-emotion's engine does not try to replace every tool a studio already owns - it plugs into them. The z-maya plugin brings cloth simulation inside Autodesk Maya. At SIGGRAPH 2024, the company showed work alongside Adobe Substance, the standard for materials and textures. The pattern is consistent: keep the physics core, meet creators inside whatever software they already live in, and let one simulation engine serve fashion, film, and games without forcing anyone to relearn their craft from scratch.
Where it points
Han's stated aspiration is democratization - making 3D fashion design affordable and intuitive enough that an individual artist, not just an enterprise studio, can pick it up without learning a different specialized program for every medium. He has carried that message to the industry's gatherings, presenting z-emotion's technology at SIGGRAPH 2024 in Denver alongside its integrations with Autodesk Maya and Adobe Substance.
It is a tidy arc for someone who started by making a video game character's hair move correctly. The threads got thicker, the surfaces got bigger, the clients started wearing suits. The physics stayed the same.