How to shut in an oil well… with machine learning. Part 2, technical

▤ Summary

This article explains how to use machine learning to predict and adjust oil well shutdown volume with custom loss functions and model workflows. It compares XGBoost and CatBoost, then shows how to tune parameters, reduce leakage, and evaluate models with composite metrics. It also describes how to build a small Flask interface for model usage and how to automate training and tracking with Airflow and MLflow. The piece is technical and implementation-focused, centered on ML engineering rather than a company announcement.

▦ Linked Companies