This document provides an overview of data science from an implementation perspective. It discusses the different mindsets of scientists, engineers, and mathematicians. Scientists are comfortable with uncertainty while engineers prefer determinism. Models are used to represent abstractions while prototypes are made for implementation. Both deterministic and probabilistic approaches are discussed. Accuracy metrics like precision and recall are important for evaluating models. The CRISP-DM process involves business goals, data preparation, modeling, and production implementation. Effective communication between product management, data science, engineering, and QA is important for success.