Published multiple technical blog posts covering a wide range of topics from theory to practice, including but not limited to distributed training techniques, application of loss functions in machine learning, and how to deploy Flask+MongoDB+Nginx environments using Docker.
Research Experience
Involved in several projects related to natural language processing, such as using Bert for sentiment classification experiments; conducted research on deep association reasoning models in community Q&A; explored deep recommendation systems based on review texts; and implemented a Pytorch version of the Factorization Machine algorithm.
Background
Research interests include but are not limited to: distributed training (PyTorch DDP), natural language processing models (such as Bert, ELMo), and applications of recommendation systems combined with knowledge graphs.
Miscellany
Enjoys listening to music; uses programming to solve practical problems, like improving program efficiency using Python's futures module; shared life reflections and personal insights from 2018 to 2020.