Responsibilities
1. Responsible for the research and development and application of Douyin content understanding algorithm product functions and basic content feature system, including basic technical support and implementation of CV, NLP, audio and other technologies in short video, live broadcast, comment and other scenarios 2. The main work directions include: (1) Responsible for the content category and intent recognition construction of short video, live broadcast, comment and other scenarios (2) Responsible for the content information summary and key information analysis of short video, live broadcast, comment and other scenarios, and the construction of industry knowledge information (3) Responsible for the research and development and application of multimodal technology, and improving the specific performance of multimodal models and LLM in business.
Qualifications
1. Have a solid foundation in machine learning, have an in-depth understanding of CV/multimodal technology, NLP technology, and GPT technology, have a solid foundation in mathematics and physics, and have strong self-study ability 2. Have solid coding skills and be proficient in using relevant machine learning frameworks and engineering frameworks 3. Have experience in the field of multimodal large models, and have experience in short video and graphic algorithms are preferred 4. Those who have published papers in high-level computer science conferences and journals such as NIPS, ICML, CVPR, ICCV, ECCV, IJCAI, AAAI, KDD, SIGIR, ACL, PAMI, IJCV, etc. or have competition experience are preferred.