You will play a central role in developing a foundational 3D generative model, advancing the state of the art in AI-generated 3D content. Your work will span the full lifecycle of model development—from pre-training and post-training to emerging techniques in diffusion and flow matching—while ensuring research translates into practical tools for creators.
Key Responsibilities
- Design, train, and optimize large-scale 3D generative models using modern paradigms including diffusion, flow matching, and multi-modal learning.
- Deploy models into production environments used by millions, incorporating human feedback and creative evaluation to refine outputs.
- Invent new architectures that improve speed, fidelity, and control in 3D generation.
- Partner with infrastructure teams to build scalable training systems and data pipelines across GPU clusters and cloud platforms.
- Apply engineering rigor in a fast-moving research setting—writing clean, maintainable code and ensuring reproducible experiments.
- Share knowledge through internal demonstrations, open-source contributions, or technical documentation to advance the broader field.
Qualifications
We’re looking for engineers with strong programming skills in Python and deep learning frameworks, particularly PyTorch, who can fluidly transition between prototyping and production. You should be familiar with Transformers and modern generative models such as VAEs, diffusion, and flow-based methods. A deep curiosity about multi-modal AI and an intuitive grasp of how models interpret and generate 3D environments are essential.
Preferred experience includes large-scale training systems like SLURM, Ray, or Kubernetes. Contributions to open-source ML projects or publications at top-tier conferences are a strong plus.
Technology Environment
Our stack includes Python, PyTorch, Transformers, diffusion and flow matching models, VAEs, and infrastructure tools including SLURM, Ray, Kubernetes, GPU clusters, and cloud environments. Work is supported through Jupyter notebooks and collaborative development workflows.
Work Environment & Culture
This is a hybrid role with an on-site office; fully remote options are available for exceptional candidates. On-site expectations vary by team. You’ll join a global team of AI researchers, computer graphics experts, and product builders in a culture that values intelligence, empathy, direct communication, and trust.
We encourage bold experimentation and prioritize quality and aesthetics across products, processes, and code. Our work is used by leading enterprises in gaming, film, robotics, 3D printing, and AI training. We believe computing’s next frontier lies in human expression and creative joy.
