Machine Learning Engineer — Customer Churn Prediction Model

We have 3 years of customer usage data in BigQuery (2.1M rows) and want to build a churn prediction model to identify at-risk customers 45 days before churn. You will explore and clean the dataset, engineer features from usage events, train and evaluate multiple models (XGBoost, LightGBM, and a neural baseline), tune hyperparameters with cross-validation, build an inference pipeline for daily batch scoring, and deploy to GCP with a simple API endpoint for our CRM to call. Target: AUC-ROC above 0.82. Deliverables include a Jupyter notebook with full analysis and a deployed model endpoint.