randxie/kaggle-pipeline

👋 Hi, I’m @randxie, a data scientist turned machine learning system engineer, currently working on Large Language Model.

Previously, I led the ML platform team in Robinhood to build feature store, ML model training / serving / monitoring and generative AI infrastructure. I also spent 2.5 years in Google Cloud - Vertex AI doing similar things (built ML infrastructure products on Google cloud). In my free time, I enjoy building open-source software and participating in data science challenges (Kaggle profile).

Before working on machine learning, I worked on machines. I hold a bachelor and master in mechanical engineering, and had spent 4 years in the field of predictive maintenance, including:

• Core developer of ML training and serving infrastructure for predictive maintenance in Uptake
• Core developer of predictive maintenance toolbox
• Student researcher working on industry 4.0 in the Center for Intelligent Maintenance System

kaggle-pipeline's Issues

Can the pipeline code be more general?

Hello, I'm a new programmer and very intersted in machine learning. I saw the your video in bittigger. What I want to know is that whether the pipeline code can be more general on feature engineering, which can handle number, image, signal and so on automatically.

fearture_importances_

Hello, I have one question about kaggle competition. Some one told me a trick: through using fearture_importances_ of tree model , we can remove the features whose importance are weak or add compound features using the important, then ,we can train model again. All of these can get better performance. I tried these tricks, but the performance was not better. Could you tell me Why this happened? Thank you very much!!!