The AI Engineering Field Guide

Ask ten people what an AI engineer does and you'll get ten different answers. Most career content about this role is built on opinions. Someone plays with a few prompts, writes a post about "the AI engineer role", and everyone else repeats it. After a while it becomes hard to tell what companies want from what people assume they want.

I got tired of that, so I built this guide on data instead. I analyzed 6,964 AI engineer job descriptions from major tech hubs, plus a large collection of interview reports from real candidates. The job descriptions show what companies ask for in their postings. The interview reports show what happens once you apply. Together they give a more honest picture of the role than another hot take.

The series walks the same path you would: understand the role, pass the interviews, then handle the offer. Each chapter is a short, practical read, and each one links back to the underlying data so you can check my numbers.

The audience for this guide

You'll get the most out of it if you're moving toward AI engineering from somewhere else. Data science, ML engineering, backend and data engineering, and general software development each get you part of the way there. I'll point out what to add for each of them.

If you're starting from zero, the guide still works. Pair it with a learning path, and I link one below, so the role description turns into a study plan instead of staying theory.

The data behind the series

The dataset behind this series covers 6,964 job descriptions scraped from a large job board across major tech hubs. Interview reports and practitioner stories complete the picture. Everything lives in a public GitHub repo: the AI Engineering Field Guide.

I keep re-scraping job descriptions and updating the analysis there, so when the market shifts, the numbers shift with it. These chapters are adapted text with my commentary, but the full breakdowns and the raw data stay in the repo. When you want the latest numbers, check the repo, not this page.

The chapters

The series has three chapters. The first two are open to everyone, and the final chapter is for members.

• Chapter 1: What Does an AI Engineer Do? - what the role is, what the day-to-day work looks like, and how it differs from data science and ML engineering. Tier: Open
• Chapter 2: The AI Engineering Interview Process - how companies structure AI engineering hiring, reconstructed from the 51 companies that publish their process and cross-checked with candidate reports. Tier: Open
• Chapter 3: After the Interview: Offers and Negotiation - what happens when the process ends: reading an offer, negotiating salary, and dealing with rejections. Tier: Main (free for Main-tier members)

Chapters 1 and 2 are open to everyone. Chapter 3 unlocks with a Main membership.

Using this guide

Start with Chapter 1 even if you think you know the role. The job is broader than prompting, and knowing where the real work sits helps you prepare for the right things. It also helps you read job postings with a critical eye. Once you know which responsibilities dominate, you can tell a serious AI role from a rebranded one.

If you're applying for jobs now, go straight to Chapter 2 after that. Interview processes for AI engineers follow patterns, and knowing the patterns removes most of the stress. You know how many rounds to expect, what each round tests, and where candidates usually stumble.

If an offer is already on your desk, Chapter 3 is for you. Negotiation feels uncomfortable for most engineers, and a bit of data makes it easier to have that conversation with a recruiter.

Related reading

Three articles on this site pair well with the series:

• The AI Engineering Job Market - the live data page behind this series, with skill demand, month-over-month trends, top employers, and locations from every monthly scrape.
• What Is an AI Engineer Based on Job Descriptions? - a deep dive into role types, skills, tools, and use cases from job posting data. If Chapter 1 leaves you wanting the full skill breakdown with percentages, this is the natural next step.
• The AI Engineer Learning Path - what to learn and in what order, with specific notes for transitioning from data science, ML engineering, and other backgrounds. Read it after Chapter 1 to turn the role description into a concrete plan.

Build this

You can read about the role in an afternoon. Building it takes longer, and that's where the learning happens.

The fastest way to start is to pick work that looks like the job:

• RAG systems over real data
• agents with tools and evaluation
• services deployed with monitoring in place

You can find practice projects designed around all three patterns on the projects page.

When you're ready for the rest of the series, become a member. A membership unlocks the negotiation chapter and the other gated content on the site.