Research Areas:

• Architecture of Large Multimodal Model (LMM), which might include VLM, World Model, Visual Foundation Model, etc.
• Unification of Complexation and Generation
• Self-supervised learning in LMM
• Few shot learning in LMM
• Training Strategies in LMM, such as RL
• Efficient Data Generation
• Automatic data processing algorithms and systems
• Systems for large-scale distributed model training
• Spatial Intelligence, such as Autonomous Driving and Embodied AI

Responsibilities:

• Being responsible for keeping track of the academic and industrial progress in LMM.
• Develop the advanced high-performance LMM architecting and algorithm to help to enhance the core competitiveness and user experience of the company's related products.
• Collaborating with product groups in development of above applications and research activities

Requirements:

• PhD/Master degree in computer vision, machine learning, computer science, mathematics, automation, or have equivalent research experiences
• Have more than two years experiences in research in applying machine learning and deep learning
• Have solid programming skills, proficient in C/C++, Python and other programming languages
• Have good team cooperation spirit, as well as communication and coordination ability
• Having the following experience will be an advantage
• Have LMM research and engineering experience
• Have experience in world model
• Have experience in training LM with RL