Andrew (Drew) Jaegle
I'm Head of AI at Mirage in NYC. We're enabling people to tell stories in powerful new ways by making compelling, expressive video creation accessible through generative AI.
Before joining Mirage (formerly Captions), I worked as a Research Scientist at (Google) DeepMind, where I developed the Perceiver family of architectures, generative models for music and creative tools (Lyria, Dream Track, Music AI Sandbox), and generative models for behavior and physical systems.
I worked on computer vision and computational neuroscience at the University of Pennsylvania, where I did a PhD in the GRASP Lab with Kostas Daniilidis and a postdoc with Nicole Rust. While at Penn, I had the privilege to spend time in the labs of Diego Contreras and Maria Geffen and with the Computational Neuroscience Initiative. During my PhD, I spent two wonderful summers at the Max-Planck Institute for Intelligent Systems, working with Michael Black and Javier Romero. I worked on consciousness and attention at the City University of New York with Tony Ro and studied philosophy, music composition, and mathematics at Texas A&M University.
Publications
Check out my Google Scholar for a full list of publications.
Talks & Media
- Justin Tranter x Music AI | Google Lab Sessions — Artist incubator with Grammy-nominated songwriter
- Embracing Creativity Through Music & AI | Inside Google DeepMind — Drew's story
- Stanford CS25: DeepMind's Perceiver and Perceiver IO — Stanford seminar
- Transformers at Work — Talk on practical transformer applications
- Perceivers: Towards General-Purpose Neural Network Architectures — London ML Meetup
Blog Posts
- Seeing Voices: Generative Video at Mirage — Mirage Blog
- Shaping Voices: Generative Audio at Mirage — Mirage Blog
- Transforming the Future of Music Creation — DeepMind Blog on Lyria
- AI and Music: The Dream Track Experiment — YouTube Blog
- Music AI Sandbox — DeepMind Blog
- Perceiver AR: Long-Context Autoregressive Generation — DeepMind Blog
- Perceiver AR for Music Generation — Google Magenta
- Building Architectures That Can Handle the World's Data — DeepMind Blog on Perceiver IO