Utopai Studios

About the Company

Utopai Studios (formerly Cybever) is an AI-native film and television studio that gives creators the power to tell epic stories as they were meant to be told, transforming finite resources into infinite worlds. Headquartered in Sunnyvale and West Hollywood, California, with a European subsidiary in Berlin, the studio develops and produces its own slate of high-profile projects while partnering with leading entertainment industry players to bring bold storytelling visions to life. Founded out of Cybever, the AI innovator behind breakthroughs in 3D world creation, procedural content generation, and real-time video generation, Utopai carries forward a legacy of technological innovation into its new role as a Hollywood production studio.

Listed Jobs

Company Name

Utopai Studios

Job Title

AI Research Scientist, Video Generation

Job Description

Job Title: AI Research Scientist – Video Generation Role Summary: Lead the design, training, and deployment of large‑scale image and video generative models. Advance continuous pre‑training, RL‑based post‑training, and unified text‑image‑video auto‑regressive architectures to enhance model performance, controllability, and creative capabilities for professional filmmaking workflows. Expectations: - PhD in Computer Science, AI, or related field (or equivalent senior research experience). - Proven industrial experience training large‑scale image/video generative models. - Strong track record of scholarly publications in top AI conferences. Key Responsibilities: - Oversee training and fine‑tuning pipelines on GPU clusters comprising hundreds to thousands of GPUs. - Develop advanced training strategies: continuous learning, reinforcement‑learning post‑training, and unified multimodal modeling. - Construct and optimize data pipelines for high‑quality multimodal datasets. - Build automated evaluation frameworks to assess fidelity, coherence, and adherence to input constraints. - Collaborate across research, engineering, and creative teams to translate breakthroughs into production workflows. Required Skills: - Deep expertise in diffusion, transformer, and multimodal generation models (text‑to‑video, image‑to‑video, video‑to‑video). - Proficiency with PyTorch, distributed training frameworks (e.g., NCCL, Horovod, DeepSpeed). - Experience with continuous training, post‑training methods, or reinforcement learning. - Strong programming, debugging, and performance profiling skills. - Ability to communicate research findings and technical concepts to cross‑functional teams. Required Education & Certifications: - PhD in Computer Science, Artificial Intelligence, or closely related discipline (or equivalent). - Published research in leading AI venues (NeurIPS, CVPR, ICCV, SIGGRAPH, etc.) is essential.

Sunnyvale, United states

Hybrid

09-10-2025