Mingbao Lin   林明宝

Working Experiences

• 2024 -- Now: Research Scientist, Kunlun 2050 Research | Skywork AI led by Shuicheng Yan, Singapore
  Low-latency Multimodal Interaction System (text, audio, image, video, etc.)
  Foundation Model Training、AI Infra、3D Reconstruction
  
• 2024 -- Now: Industry Supervisor, AI Institute | Xiamen University led by Rongrong Ji, Xiamen, China
  Multimodal Model (MLLM, Diffusion) for multimodal tasks (Training-Free/Tuning-Efficient Editting)
  Model Compression/Inference Acceleration/Training Efficiency on CV (ViT, etc.), NLP (LLM etc.) and Multimodal (MLLM, Diffusion etc.) models.
• 2022 -- 2024: Senior Researcher, Youtu Lab | Tencent, Shanghai, China
  Large Language Model (LLM) for general NLP tasks (Foundation Model Training, Downstream Applications)
  Model Compression on CV (CNN, ViT, etc.) and NLP (Bert, etc.) models for general tasks (Recognition, OCR, etc.)
• 2019 -- 2022: Research Intern, Peng Cheng Lab, Shenzhen, China
  Model Compression (Quantization, Pruning, Sparsity, Distillation, etc.) on general CV models (CNN, ViT)
  Low-level CV (Super-Resolution, Shadow Removal, Demoiréing, etc.)

Education Experiences

• 2016 -- 2022: M.S.-Ph.D. in Comput. Sci., MAC Lab led by Rongrong Ji, Xiamen University, China
  Model Compression (Quantization, Pruning, Sparsity, Distillation, etc.) on general CV models (CNN, ViT)
  Low-level CV (Super-Resolution, Shadow Removal, Demoiréing, etc.)
  Image/text retrieval
• 2012 -- 2016: B.S. in Comput. Sci., Fuzhou University, Fuzhou, China