NYC Short-Term Rental Price MLOps Pipeline

A reproducible machine-learning pipeline for estimating nightly NYC short-term rental prices, designed around clean handoffs from raw data to a promoted model artifact. MLflow orchestrates ingestion, cleaning, validation, splitting, training, and regression testing; Hydra manages configuration; and Weights & Biases provides dataset lineage, experiment tracking, drift references, and model lifecycle aliases. A Random Forest combining structured features with TF-IDF listing-title signals achieved a held-out MAE of 33.29 and R² of 0.581, backed by geographic checks, price-range gates, and KL-divergence drift detection.