Image credit: From left to right: Cassidy Curtis, Meera Hahn, Ellen Jiang, Ben Wedin, Andreas Veit
In this episode of SIGGRAPH Spotlight, Production Sessions Subcommittee member Nathaniel Peters is joined by members of the “Dear Upstairs Neighbors” team to explore the role of generative AI in animation. Drawing from their collaborative work at Google DeepMind, the conversation examines how artists, researchers, and engineers came together to integrate emerging tools into a production pipeline. They also highlight the balance between experimentation and control, the realities of working with AI in practice, and how these evolving technologies are reshaping creative workflows, storytelling, and collaboration across disciplines.
Don’t miss a single episode by subscribing to the show on Apple Podcasts (iTunes), Spotify, or SoundCloud. If you like the show, rate and review wherever you listen to help other listeners find us!
About Our Guests
Cassidy Curtis (supervising animator for Dear Upstairs Neighbors) grew up with computer animation (his first SIGGRAPH was 1989) and he’s been practicing it ever since, in many different roles: developing effects animation software at Xaos Inc, doing non-photorealistic rendering research at the University of Washington, animating characters at PDI/DreamWorks, making real-time immersive films at Google Spotlight Stories, and combining all of these interests with machine learning at Google DeepMind. He has the unusual distinction of having won VES awards for his work both as a character animator (Toothless in How to Train Your Dragon) and look developer (Age of Sail). His mission is to help artists and technologists understand each other, and do things together that they couldn’t possibly do alone.
Meera Hahn is a research scientist at Google DeepMind, working on bridging the gap between complex AI models and pure entertainment. She started her career focused on the theoretical and technical aspects of machine learning, and over time, shifted her focus to the frontier of generative and creative AI. This has led to her work as a technical artist, where she has helped bring to life cutting-edge creative projects like Wizard of Oz at Sphere and Dear Upstairs Neighbors. Meera’s goal is to build AI models that don’t replace human artists, but instead amplify their creativity and unlock entirely new forms of entertainment.
Ellen Jiang is a research engineer at Google DeepMind, where she designs new ways for people to create and work with AI. She’s also a practicing artist — working in digital illustration and gouache painting — and previously interned in computer animation as part of Pixar’s Undergraduate Program. Dear Upstairs Neighbors was a rare chance to bring these worlds together, collaborating with researchers and artists to pioneer a new kind of filmmaking pipeline.
Ben Wedin is a Research Engineer at Google DeepMind. As part of the People + AI Research (PAIR) team, he helps research and design systems to make AI partnerships productive, enjoyable, and fair.
Andreas Veit is a Research Scientist at Google DeepMind focusing on Machine Learning and Computer Vision. He began his career focusing on image and video understanding, spending several years advancing search and discovery technologies for platforms like Google Search, YouTube and Maps. In recent years, Andreas has shifted his focus to generative AI. As someone with an artistic spirit who felt he previously lacked the traditional mechanical skills to fully express his ideas, he finds generative AI to be a deeply empowering medium.
