NYTech
People
Personfounder

Anastasis Germanidis

Anastasis Germanidis is a computer scientist known as the co-founder and chief technology officer of Runway, a New York based company that develops generative image and video models for creative applications [1][2]. He holds a bachelor's degree in computer science from Wesleyan University and a master's degree from New York University's Interactive Telecommunications Program, an art and technology program where he met his future Runway co-founders [10][11][13]. Before co-founding Runway in 2018, he worked in backend and machine learning engineering roles, including as a backend engineer at Quantcast and Chartbeat, a computer vision researcher at IBM Research, and briefly as a backend engineer at Zocdoc [6][7][8][9]. He has described his background as a combination of computer science work and long-standing interest in applying machine learning, particularly convolutional neural networks, to artistic and creative problems, tracing this interest back to reading about neural networks in high school [13].

Germanidis has described Runway as an "applied research company" that conducts research into new AI techniques and then deploys them as tools for creative teams [14]. According to his account, the company originated from efforts by him and his co-founders at NYU to make existing generative models, including Pix2Pix, GANs, and style-transfer methods, easier for artists and filmmakers to use, since deploying such models at the time typically required extensive technical setup [13]. He has argued that each time Runway has simplified access to a model, usage among non-technical creatives expanded, a pattern he says has held from Runway's early open-source tools through the release of its Gen-1 and Gen-2 models [13]. He has also stated that Runway trains models on the order of tens per week as part of this research process [13].

He has described Gen-2, Runway's text-to-video model, as the outcome of years of prior research the company conducted on image generation, with techniques and architectures from that work later adapted and retrained on large sets of video sequences to allow the resulting models to learn temporal dynamics [13][14]. He has said that Runway's tools, including a rotoscoping tool for isolating subjects from video backgrounds and an inpainting tool for localized image editing, are used across pre-production, production, and post-production stages of video work, and has cited use of Runway's visual effects tools on the film Everything Everywhere All at Once [13][14]. He has also pointed to teams at CBS, including graphics staff for The Late Show with Stephen Colbert, as examples of professional users who rely on Runway's tools to produce content under tight turnaround times [13].

Germanidis has stated that Runway offers both a direct consumer product, allowing individuals to sign up and use the tools without an enterprise agreement, and enterprise plans for larger organizations that need collaboration features or custom-trained models built on their own image or video datasets [14]. He has said Runway's longer-term aim is to move toward generating narratively coherent, feature-length video output incorporating visual, sound, and dialogue elements [13]. He has framed the technology's role as expanding the range of ideas creators can attempt and speeding iteration, while stating that the tools serve as a multiplier of a person's ability to realize a creative vision rather than as a substitute for that vision [14].

Founded

Experience

  1. Co-founder & Co-CEO
    RunwayFeb 2026 to Present
  2. Co-Founder & CTO
    RunwayNov 2018 to Feb 2026
  3. Backend Engineer
    ZocdocJun 2018 to Nov 2018
  4. Machine Learning Researcher - Computer Vision
    IBM ResearchJun 2017 to Jan 2018
  5. Backend Engineer
    ChartbeatMar 2015 to Sep 2016
  6. Backend Engineer
    QuantcastJun 2013 to Jan 2014

Education

Media & appearances

In the news

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