← Back to all free guides
A free guide by Mariah Brunner ↗ INSTAGRAM ↗ TIKTOK

The $200K Job
Died. This One
Replaced It.

Everyone rushed to learn prompt engineering, and the job title already faded. The skill that’s sticking is context engineering. Here’s what it is and how to start building it today.

In 2023, prompt engineering was supposed to be THE job of the future, with talk of $200,000 salaries for writing good prompts. By 2025, Fortune called the role obsolete. Job searches for it peaked in April 2023 and kept falling.

The skill didn’t stop mattering. Indeed’s economist put it this way: "Prompt engineering as a skill is still definitely a good thing to have, but it’s not an entire title." Everyone is now expected to know it.

What Context Engineering Is

Context engineering covers everything the model sees before it answers, from your instructions and documents to the chat history and tools. Anthropic defines it as "the set of strategies for curating and maintaining the optimal set of tokens (information)" the model works with. (Anthropic)

The Simplest Way To Think About It

A prompt engineer asks a better question. A context engineer builds a better desk for the AI to work at, with the right files open, the right rules pinned up and the right tools within reach. Same model, much better output.

01

What it looks like

Prompt only: "Write a follow-up email to a customer who didn’t reply."

With context: You paste in the original thread plus your brand voice guide. You add the customer’s order history too. Then you ask for a follow-up in your voice that mentions their order and gives one clear reason to reply this week.

The second one wins because the AI had what it needed to answer well.

The 6 Parts You Control

"Everything the model sees" is too vague to act on. These are the six parts. Almost every bad AI answer is missing one of them.

  1. Instructions. Who the AI is acting as, plus your do’s and don’ts.
  2. Knowledge. Your actual docs, data and numbers, pasted in. A description of them is far weaker.
  3. Examples. One or two samples of a great answer. Showing beats describing.
  4. History. What to carry forward and what to drop. Old, off-topic chat drags answers down.
  5. Tools. What the AI can reach, like your files, the web, your calendar or your inbox.
  6. Format. How you want the answer back, said up front.
Steal My Context Template

Keep this in a note and paste it at the top of any task that matters. It forces all six parts into place. Notice the actual ask goes last.

The Context Template
ROLE: You are my [job title or expert, e.g. senior email marketer]. GOAL: I need [the exact outcome, e.g. a 3-email win-back sequence for customers who haven't bought in 90 days]. CONTEXT (the real material, pasted or attached): <documents> [paste the actual doc, data, thread or numbers here] </documents> - About me and my business: [paste your "how I work" file] - Who this is for: [your audience or customer] RULES: - Do: [what a great answer must include] - Don't: [what to avoid, e.g. no hype, no made-up stats] - Use only the material above for facts. If something isn't in it, say so instead of guessing. - If you're missing something you need, ask me up to 3 questions before you start. EXAMPLE OF GREAT (optional, but powerful): <example> [paste one example of the kind of output you want] </example> FORMAT: Give it to me as [a table / 5 bullets / a short doc], under [length]. TASK: [your actual ask, in one or two sentences, last]

Use it once by pasting it into a new chat. To keep it, make a Claude Project, click Set project instructions and paste in the parts that never change, like your role, rules and "how I work" file.

The Habits That Make You Good

01

Tight beats big

Anthropic calls the failure "context rot": as the context window fills up, the model gets worse at recalling what’s in it. Give it what the task needs, then stop. (Anthropic)

02

Put the ask last

With long documents, put them at the top and your question at the end. Anthropic says this can improve answer quality by up to 30% in its tests. (Claude docs)

03

Change the context, not the question

When an answer is off, don’t just reword the question. Add the missing piece and run it again. That one habit is the skill.

04

Start fresh often

When a chat gets long and cluttered, open a clean one and paste a tight context block.

3 Drills For This Week
  1. The rebuild. Take a weak AI answer you got today. Run it through the coach prompt below, add the one missing piece, and run it again.
  2. The golden file. Write your "how I work" file with your role, goals, voice and rules. Refine it every time you catch yourself re-explaining something.
  3. The trim. Take a long chat and cut the context in half. If the answer still holds, you just learned what the model never needed.
The Missing Context Coach
You are my context engineering coach. I'm going to give you a task I used AI for and the answer I got back. Your job is to find what context was missing, not to rewrite the question. Here's what I asked: [paste your original prompt] Here's what I got: [paste the AI's answer] Here's what I actually wanted: [one or two sentences] Steps: 1. Check my prompt against these six parts: instructions (role and rules), knowledge (real docs or data), examples, history (what should carry over or be dropped), tools (files, web, connectors it should use), and format. 2. For each part, mark it Present, Weak or Missing, with one line on why. 3. Pick the ONE missing piece that would fix the answer most, and tell me exactly what to paste or add. 4. Write the improved prompt using my template order: role, goal, context, rules, example, format, then the task last. Use [brackets] for anything only I can fill in. Keep it short. Don't invent facts about my business. If you can't tell what I wanted, ask me before you write anything.

Jensen Huang put it plainly: "It is most likely that most people will lose their job to somebody who uses AI." (Fortune) Context engineering is the durable version of that skill.

Where To Learn More
  1. Anthropic: Effective Context Engineering for AI Agents. The clearest breakdown, from the team that builds Claude. Start here.
  2. Andrej Karpathy’s post. He helped popularize the term with "+1 for ‘context engineering’ over ‘prompt engineering’."
  3. Prompting Guide: Context Engineering. A free walkthrough with examples, great for practice.
  4. Claude Cookbook: Context Engineering. Hands-on, for when you start building agents.

If you do one thing today, read the Anthropic piece and run drill one on a real task from your job.