Machine Learning Engineer from Dalian, Liaoning, CHN, China
Masters in Civil Engineering, research mainly in distribution of heat and energy in buildings. Published a paper Anomaly detection using convolutional autoencoder with residual gated recurrent unit...
Masters in Civil Engineering, research mainly in distribution of heat and energy in buildings. Published a paper Anomaly detection using convolutional autoencoder with residual gated recurrent unit and weak supervision for photovoltaic thermal heat pump system https://www.sciencedirect.com/science/article/abs/pii/S2352710224032625
Been doing machine learning for at least 4 years, and participated in many projects, including kaggle competitions, United Nations Education Programme on Environmental sustainability.
Worked as a data engineer for a chinese startup, dealing with web scraping, crawling, annotation (use multiple llms with confidence voting), sentiment analysis (Chinese-Bert and Embeddings for longer content since bert is limited to 512 tokens), local deployment using docker and nginx
Currently working as AI engineer for another Chinese startup, this time, working on automating the review of architectural drawings for compliance using AI. The main tasks involve transforming CAD to structured data using ezdxf for extracting data, shapely and pydantic for creating well structured data, also convert DXF files to images in order for LLMs to have visual understanding rather than just vector data. Tried different frameworks, but dspy won my heart, lately, Agentscope by Alibaba, the team behind the popular models Qwen has been a delight, native support for mcp, thus improved workflow for the agent since I can reuse the work of other people.