This document provides an overview of schema design patterns for MongoDB databases. It discusses common issues like large documents, working sets not fitting in RAM, and repeated calculations. Specific patterns covered include Attribute, Subset, Computed, and Approximation. The Attribute pattern addresses issues with similar fields across documents by storing them in an array. The Subset pattern embeds a small subset of fields to reduce document size. The Computed pattern pre-calculates values to avoid repeated computations. Approximation stores approximate values for non-critical data to reduce writes. The talk encourages using these patterns as building blocks to design schemas for performance, scalability and simplicity.