Andrés Romero

I am a machine learning and computer vision expert with over 9 years of experience, currently serving as Head of AI at Haxion, where I lead the development of cutting-edge generative AI products. I specialize in Diffusion Models, Large Language Models, and Neural Rendering, with proven expertise in 3D human avatars, computational photography, and synthetic data generation.

At Haxion, I direct a team of 9 engineers developing two products: a controllable image/video generation platform using advanced diffusion models that provides users with precise creative control, and a camera-based 3D editing system that enables natural language interaction with 3D scenes through point clouds and Gaussian splatting technology.

My research background includes pioneering work in facial expression generation using GANs during my PhD, followed by postdoctoral research at ETH Zürich focusing on image synthesis and video editing. I have successfully bridged academic research with practical applications, contributing to open-source projects and organizing major computer vision challenges at top-tier conferences.