Video ML Engineer (Remote — Worldwide)

Join the world-class research and engineering team at Runway to push the boundaries of what is possible with generative video. As a Video ML Engineer, you will design, train, and deploy large-scale diffusion and transformer models that power products like Gen-2 and beyond. You will work at the intersection of high-performance computing and creative expression, directly impacting how millions of creators tell stories through AI.

- Design and implement state-of-the-art architectures for high-fidelity video synthesis and editing. - Optimize large-scale training runs on multi-node GPU clusters to improve model efficiency and sample quality. - Develop novel techniques for temporal modeling and long-form video consistency. - Collaborate with the product team to integrate research breakthroughs into the Runway web and mobile interfaces. - Experiment with multi-modal inputs including text-to-video, image-to-video, and video-to-video workflows. - Stay current with the latest literature in generative AI and contribute to the internal codebase and research library. - Evaluate and benchmark model performance using both automated metrics and human-in-the-loop feedback.

- Master’s or PhD in Computer Science, Machine Learning, or a related field with a focus on Computer Vision. - Proven experience training large-scale generative models (GANs, Diffusion, or Autoregressive Transformers). - Strong proficiency in PyTorch and experience with distributed training frameworks like DeepSpeed or FSDP. - Deep understanding of video temporal consistency, optical flow, and 3D architectural priors in ML models. - Experience managing large-scale datasets and implementing efficient data-loading pipelines for video. - Track record of shipping production-grade ML models or publishing at top-tier conferences (CVPR, ICCV, NeurIPS). - Ability to work independently in a remote, fast-paced research environment across global time zones.

Equity, benefits, remote.