Responsibilities
Team Introduction: PixelDance is a video generation model developed by ByteDance Research, which can generate up to 10 seconds of exciting video clips at one time. The model has accurate semantic understanding ability, powerful dynamic effects and cool camera movements, can achieve consistent multi-lens generation, and is compatible with multiple styles and proportions. ByteDance Research is committed to cutting-edge technology research in the field of artificial intelligence, covering natural language processing, computer vision, machine learning, robotics, AI for Science, Responsible AI and other fields. At the same time, it will implement research results and provide technical support and services for the company's existing products and businesses. 1. Deeply participate in the product work of the cutting-edge direction of large models for video generation, and work closely with algorithm, engineering and other teams to create first-class technology and products 2. Driven by user scenarios, guide the algorithm team to focus on important directions, achieve first-class results in the industry, and succeed in important businesses within the company 3. Responsible for large model strategies, define the ideal state of models and data, continuously complete project breakthroughs, greatly improve model effects, and be responsible for the construction of a scientific and objective evaluation system 4. Explore the application and implementation of cutting-edge video generation in actual user scenarios and needs.
Qualifications
1. Bachelor degree or above, more than 3 years of AI product experience, computer science related majors are preferred, and those with experience in video generation, text and graph basic model strategy products are preferred 2. Have strong analysis and communication skills, good at discovering valuable product improvement suggestions from data and evaluation, and can promote implementation 3. Have a certain understanding of algorithm principles, familiar with technical principles such as Stable Diffusion, LLM, Sora, and have the ability to read papers. 4. Have enthusiasm and love for GenAI, have excellent curiosity and learning ability, can adapt to a fast-paced working style, are willing to accept high-challenging goals, and can continuously tackle tough problems and achieve results.
