1. First Principles of Machine Learning Engineering

Before diving into specific AWS services and exam domains, this phase builds the mental models you'll use to reason through every question on the MLA-C01. You'll learn why ML engineering is fundamentally different from traditional software engineering, how the four-stage ML lifecycle (Data → Model → Deploy → Monitor) maps onto AWS services, and how the SageMaker ecosystem ties the stages together. Most importantly, you'll internalize the cost-performance-latency trade-offs that underpin nearly every scenario-based question on the exam.

None of the content in Phase 1 maps directly to a single exam domain — it spans all four. Think of it as the operating system that Phases 2–5 run on. When a Phase 3 question asks you to choose between XGBoost and a foundation model, you'll reach for the "simplest model that solves the problem" principle from 1.4. When a Phase 5 question describes declining accuracy, you'll recognize the lifecycle loop from 1.2. Build these foundations well and the domain-specific content clicks into place.