Zijian He

Zijian He is the Chief Scientific Officer at Utopai Studios, an AI-native entertainment company pioneering a new studio system for the AI era. By bridging proprietary AI infrastructure, original IP development, and deep production expertise, Utopai Studios champions a human-led model to develop, produce, and scale long-form cinematic storytelling across film, television, animation, and digital platforms.

‍
At Utopai Studios, He leads the company’s AI research efforts, including the development of next-generation multimodal foundation models for cinematic creation. His work is central to PAI,
Utopai Studios’ cinematic storytelling AI system and production intelligence layer, which is being built to support the realities of professional entertainment production, including long-form
narrative structure, visual continuity, scene planning, character and world consistency, and iterative creative development.

‍
He’s research helps advance Utopai Studios’ broader mission of defining the infrastructure layer for AI filmmaking. Rather than treating AI as a standalone generation tool, Utopai is building systems that integrate intelligence directly into the creative and production pipeline, giving filmmakers, studios and global media partners greater control, speed and scalability while keeping human creativity at the center of the process.

‍
Before joining Utopai Studios, He led the Media Generation foundation model team at Meta’s Superintelligence Lab and co-led major generative AI initiatives, including Meta Movie Gen and Llama Image Generation. Earlier in his career at Meta, he led teams developing foundational AI technologies deployed across consumer-scale products, including real-time depth estimation systems for Quest 2 and Quest 3, as well as on-device image recognition and detection models used across Facebook, Instagram and Messenger.

‍
With more than a decade of experience building AI systems at global scale, He brings deep technical expertise across multimodal modeling, large-scale pretraining, high-quality supervised fine-tuning, reinforcement learning, generative AI and computer vision. At Utopai Studios, he is applying that experience to one of the most complex frontiers in AI: building production-ready systems that can support ambitious, long-form storytelling without removing the human
judgment, authorship and creative intent that define great entertainment.

‍