多模态生成算法工程师/Multimodal Generation Algorithm Engineer

岗位职责

• 研发，跟踪，改进最新的3D视觉相关算法，根据具体业务场景适配并提升VGGT/MASt3R/ViPE等多模态3D算法；
• 研发，跟踪，改进视频生成基础模型相关结构的研究和验证，利用多模态先验提升视频生成的可控性和一致性；
• 研发，跟踪，改进单物体与场景级3D重建生成算法开发，包括文生3D、图生3D等内容；
• 根据产品的应用场景和需求，完善和开发3D AI算法，利用AI提升产品效率和效果。

任职要求

• 对人工智能技术充满热情和信仰，具有强烈的探索精神和创新意识；
• 计算机、应用数学、模式识别、机器学习、机器人等相关专业硕士或博士学位，或有同等工作经验；
• 熟悉深度学习框架（TensorFlow/PyTorch），深入理解生成模型（如VAE、Stable Diffusion、VAR等）；
• 精通数据结构、算法设计和并行计算，有良好的编程能力（Python/C++等）；
• 具备较强的创新能力和团队协作精神，能够在高压环境下保持冷静、推动解决复杂问题。

Job Description

• Research, track, and improve the latest 3D vision-related algorithms, adapt and enhance multimodal 3D algorithms such as VGGT/MASt3R/ViPE according to specific business scenarios;
• Research, track, and improve the research and verification of video generation foundation model-related structures, and use multimodal priors to enhance the controllability and consistency of video generation;
• Research, track, and improve the development of single-object and scene-level 3D reconstruction generation algorithms, including text-to-3D, image-to-3D, etc.;
• Improve and develop 3D AI algorithms according to the application scenarios and requirements of products, and use AI to improve product efficiency and effectiveness.

Job Requirements

• Passionate about and believe in artificial intelligence technology, with a strong spirit of exploration and innovation;
• Master's or doctoral degree in computer science, applied mathematics, pattern recognition, machine learning, robotics or related majors, or equivalent work experience;
• Familiar with deep learning frameworks (TensorFlow/PyTorch), and have an in-depth understanding of generative models (such as VAE, Stable Diffusion, VAR, etc.);
• Proficient in data structures, algorithm design and parallel computing, with good programming skills (Python/C++ etc.);
• Strong innovation ability and team collaboration spirit, able to stay calm under high pressure and promote the solution of complex problems.