#200582588_EN-1
he candidates should have deep understanding on the whole stack of 3D computer vision and graphics algorithms, knowing the limitation and can-do of the state-of-the-art algorithms in order to make accurate design choice and have a proficient communication with other algorithm teams. The candidate should also have excellent software developing skills in not only implementing the algorithm, but also in code optimization on heterogeneous platforms, so you can write performant algorithms on mobile and desktop or transferring workload to accurate hardware and chips. Since we are doing something that is not solved by others, we usually need to develop tools that have not yet existed in the world, it requires the candidate to have experience and capability in designing and building extensible and practical computer vision system or algorithm frameworks from ground up.
Minimum Qualifications
PhD/MS in 3D computer vision or machine learning; alternatively, a comparable industry career, with significant experience on delivering products using state-of-the-art computer vision and machine learning technologies.
Expertise in at least one area of computer vision and machine learning (e.g., multi-view geometry, stereo, photometric stereo, neural 3D reconstruction, differential rendering, diffusion model, vision-language model, 3D generative AI, RGB-D and LIDAR sensor fusion). Candidates with publication record in relevant conferences and journals (e.g. PAMI, IJCV, ToG, NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, SIGGRAPH) are preferred
Excellent programming skills in C++ with hands-on experiences in building a practical vision system or algorithm framework, e.g., SLAM system, 3D reconstruction system, differentiable rendering system, or machine learning framework, from ground up
Proficiency in at least one major machine learning framework, such as Tensorflow, PyTorch etc.
Preferred Qualifications