Overview

Modern software platforms — from industry leaders such as Google, Meta, and OpenAI to early-stage startups — depend fundamentally on data storage. This course is grounded in a core principle: data storage is not merely a technology, but a solution to a specific engineering problem. Understanding which problem is being solved is what drives every sound architectural decision.

Students will gain comprehensive exposure to the full spectrum of contemporary data storage paradigms, including relational databases (RDBMS), key-value stores, NoSQL, document stores, column-oriented databases, OLAP systems, vector databases, and embedded and serverless solutions. For each category, the course examines practical trade-offs, strengths, and limitations in real-world software contexts.

A central focus of the course is developing the judgement to evaluate data storage requirements within a given software system and to select the most appropriate solution — or combination of solutions — aligned with both business objectives and architectural constraints, such as monolithic or microservice designs.

Beyond technical knowledge, students will build the conceptual models necessary to reason about data storage independently: as hands-on engineers, as technical managers, or as practitioners working alongside AI coding assistants such as ChatGPT. The course also addresses the rapidly growing domain of AI-native systems, covering how to select and configure data storage for AI Agents — providing them with persistence, contextual memory, and the capacity to evolve.