NYTech
People
Personfounder

Alejandro Matamala

Alejandro Matamala Ortiz is a Venezuelan-born, Chile-raised designer and programmer who co-founded Runway and serves as its Chief Design Officer, guiding the New York company's work on generative image and video models for creative industries [1]. He traces the company's origins to New York University's Interactive Telecommunications Program around 2017, where he and his co-founders, coming from art and design backgrounds rather than pure computer science, began experimenting with early generative adversarial network research and applying it to creative workflows [4][2]. He recalls that the first image models he encountered produced crude, low-resolution outputs with poor text-image alignment, yet he and his collaborators became convinced the technology would eventually be capable of powering full-length narrative films [3][2]. That conviction led to Runway's founding in 2018 as what Matamala describes not as an AI company but as "a media and entertainment company," a framing he says reflects the founders' creative training and their goal of measuring success by the volume of stories and films the tools enable rather than by technical benchmarks alone [4][3].

Matamala describes Runway's product evolution across four generations of image and video models, noting a shift in customer behavior from 2018, when most clients treated the technology as an experimental R&D curiosity, to 2025, when organizations expect to deploy it directly in production creative pipelines [2]. He frames current research priorities around what Runway calls "general world models," an effort to move beyond pixel-level visual generation toward systems that better capture physics, complex body movement, and camera dynamics, drawing on varied data sources including sensor and synthetic data [3][2]. On the question of model architecture, he says Runway has invested heavily in diffusion-based approaches in recent years while continuing to run parallel experiments and remaining open to alternative techniques [3]. He also argues that interface design has lagged behind model capability, contending that text prompts alone are insufficient for creative control and that the company has been building additional "knobs" allowing more granular adjustment of outputs [4][2].

Describing his own design philosophy, Matamala uses the metaphor of "carving" products out of research models rather than architecting them in advance, saying that attempts to anticipate a model's capabilities before working directly with it have generally failed [2]. He says Runway addresses this by embedding design engineers alongside traditional product designers and pairing researchers with veteran visual-effects professionals so that speculative, functional prototypes can be built early against real model constraints [2]. On industry partnerships, he points to collaborations with companies including Getty and Lionsgate as part of an effort to secure rights-appropriate training data and to understand how studios and advertisers want to integrate generative tools into existing workflows, while acknowledging that norms around copyright and compensation in this area remain unsettled [3].

Founded

Insights & ideas

Alejandro Matamala frames Runway as a media and entertainment company rather than a pure AI company, tracing this identity to its origins at NYU's interactive telecommunications program around 2017, where he and his co-founders, coming from design and creative backgrounds, built generative AI tools they themselves wanted to use as artists [1]. He argues that the company's core motivation has always been empowering creators and artists to tell stories, not simply advancing AI technology for its own sake, and he measures success by the number of movies and stories people are able to produce with these tools rather than by technical benchmarks [1].

He also emphasizes a shift in creative control, arguing that generative models must move beyond simple text prompts toward more granular "creative knobs" that give artists precision and full controllability over outputs [1]. For Matamala, high fidelity or resolution alone is insufficient if creators cannot precisely tweak and direct what a model generates, and he envisions this trajectory eventually enabling fully generated feature-length films [1].

Media & appearances

This page shows public professional information only, each fact cited. Is this you? Corrections or removal within 24 hours, no questions asked.