Research Scientist, Professional Creative Workflows

Role overview

A research scientist position focused on pushing the frontier of generative AI for professional creative work. The role spans image, video, 3D, vector graphics, and design, blending ambitious long-horizon exploration with prototype systems that have real product impact. The ideal candidate can move fluidly between cutting-edge research, rigorous experimentation, and the practical concerns of how creators actually work.

Responsibilities

- Conduct original research on generative AI for professional creative workflows across image, video, 3D, vector graphics, and design modalities. - Develop new methods that advance the state of the art while remaining relevant to real creative and production pipelines. - Publish at leading academic venues where appropriate and contribute to high-impact internal projects. - Explore multimodal generative systems that integrate different media types, with emphasis on controllability, editability, and reusable asset creation. - Translate research insights into prototypes, experiments, and model capabilities that inform product direction. - Design datasets, benchmarks, and evaluation protocols that measure quality, editability, consistency, and production usefulness. - Collaborate with researchers, engineers, designers, and product teams to identify research problems with meaningful downstream impact. - Track frontier developments in generative media and identify opportunities to combine ideas across computer graphics, vision, language, and creative tooling.

Requirements

- Strong academic research track record in computer vision, computer graphics, machine learning, generative modeling, computational imaging, or visual computing. - PhD or equivalent research experience in Computer Science, Machine Learning, Computer Graphics, Computer Vision, or a closely related discipline. - Demonstrated ability to conduct original research, publish papers, and build working systems or prototypes. - Hands-on experience with generative and multimodal modeling, including diffusion models, transformers, 3D generation, neural rendering, inverse rendering, video generation, image editing, controllable generation, or design automation. - Broad technical background spanning several areas such as generative image models, video generation or editing, 3D asset generation or geometry processing, differentiable rendering, vector graphics and typography, multimodal representation learning, or professional creative tooling. - Strong programming skills in Python and PyTorch, with the ability to implement, adapt, and extend modern research code. - Ability to reason about both research quality and workflow quality, including controllability, iteration speed, editability, consistency, and integration into existing creative processes. - Strong communication skills and the ability to explain research ideas clearly to both research and product-oriented collaborators. - Evidence of creative technical judgment: combining ideas from different fields into new research directions rather than only incremental improvements.

Nice to have

- Publications at top-tier venues such as CVPR, ICCV, ECCV, SIGGRAPH, SIGGRAPH Asia, NeurIPS, ICLR, or ICML. - Research experience spanning more than one modality, especially across image, video, 3D, vector graphics, design, or multimodal generation. - Background in professional creative domains such as VFX, games, animation, product visualization, advertising, graphic design, or 3D asset production. - Experience with controllable generation, layered image editing, material and lighting decomposition, multi-view consistency, relightable assets, vector-aware generation, or structured design outputs. - Experience building research prototypes that influence product direction or enable new product capabilities.