AI Engineering Foundations

What you will be able to do

• You will be able to describe what an AI engineer does day to day and how the role differs from machine learning engineering, data science, and traditional software engineering.
• You will be able to explain the foundation-model shift and why building on pretrained models, rather than training your own, reshaped the discipline.
• You will be able to map the modern AI application stack across its model, orchestration, context and retrieval, tool, evaluation, and guardrail layers.
• You will be able to trace how a single request flows through that stack and point to where cost, latency, and failure tend to appear.
• You will be able to make a foundation-model API call and read every part of the request and response: messages, roles, parameters, tokens, and finish reasons.
• You will be able to handle the realities of a production API call, including keys, rate limits, errors, retries, streaming, structured output, and cost estimation.
• You will be able to apply a build-versus-buy framework that weighs cost, control, data and privacy, latency, capability, and long-term maintenance.
• You will be able to choose between a managed model API, self-hosted open-weights models, and an off-the-shelf product, and defend that choice against vendor lock-in.

What is inside

4 modules, 8 lessons. Each module ends in a graded quiz and most carry an assignment.

1. 01

   The AI Engineering Discipline

   Before you write any code, get clear on the job itself. This module defines what an AI engineer does day to day, how the role differs from machine learning engineering, data science, and software engineering, and why capable foundation models moved the work from training your own models to building on pretrained ones. You leave with an accurate mental model of the discipline and where you fit within it.

   2 lessons · 5 quiz questions

2. 02

   The Modern AI Application Stack

   AI products are built in layers, from the model up through orchestration, context and retrieval, tools, evaluation, and guardrails. This module walks each layer, then follows a single user request through all of them so you can see exactly where cost, latency, and failure live.

   2 lessons · 5 quiz questions · assignment

3. 03

   Using Foundation Models Through APIs

   The foundation model API is the primitive an AI engineer touches every day. This module takes one request and response apart, then hardens it for production. Lesson 1 covers the anatomy: messages and roles, the parameters that shape generation, tokens and the context window, and the finish reasons in the response. Lesson 2 covers the realities of live traffic: API keys and secrets, rate limits, the error and retry taxonomy, streaming, structured output, and how to estimate cost.

   2 lessons · 5 quiz questions

4. 04

   Choosing Between Building and Buying

   Every AI project starts with one decision that shapes everything after it: build or buy. This module gives you a six-dimension framework covering cost, control, data and privacy, latency, capability, and maintenance, then applies it to managed model APIs, self-hosted open-weights models, and off-the-shelf products, so you can weigh total cost of ownership and spot vendor lock-in before it surprises you.

   2 lessons · 5 quiz questions · assignment