Machine Learning Engineer – Customer Churn Prediction Model, Built & Deployed

We need an ML engineer to build and deploy a customer churn prediction model for an e-commerce business. You'll take raw customer data, build a model that identifies customers likely to churn, and deploy it as a live API we can integrate into our workflows. Scope of work: Phase 1 — Data & Model Development Clean and explore e-commerce customer data (purchase history, order frequency, recency, engagement signals, demographics — dataset to be shared on award) Engineer features (RFM — recency/frequency/monetary — behavioral signals, etc.) Train and evaluate churn classification models (e.g., XGBoost, Random Forest, Logistic Regression baseline) Target: identify customers likely to churn in the next 30/60 days Validate with appropriate metrics (F1, precision/recall, ROC-AUC — given churn is typically imbalanced) and document trade-offs Phase 2 — Deployment Package the trained model behind a REST API (FastAPI or Flask) Containerize with Docker Deploy to a cloud environment (Azure or AWS) Set up logging so churn predictions can be tracked and re-evaluated over time Phase 3 — Handoff Clean, documented GitHub repo Short write-up: architecture, retraining steps, redeployment steps Live walkthrough call Requirements: Proven experience with churn/classification modeling on imbalanced datasets Strong Python, scikit-learn/XGBoost, and API development skills Docker and at least one cloud platform (Azure preferred) Comfortable working independently with milestone check-ins Nice to have: Experience specifically with e-commerce or subscription churn MLflow or similar for experiment tracking Project type: Fixed-price, milestone-based (Phase 1 / Phase 2 / Phase 3) Estimated duration: 3–4 weeks To apply: Include a link to a past churn or classification project where you handled both modeling and deployment.