Context engineering designs everything a model receives during a task, whereas prompt engineering focuses on the words in one request. Context can include instructions, conversation history and summaries. It can also include source files, tests and documentation. Tool results can land there too.
A strong prompt cannot fix missing or misleading context. If the model cannot see the current interface or a project rule, it must guess. Context engineering puts the right information in front of the model at the right time, and removes noise that could distract the model.
Context engineering also decides how context is retrieved, ordered and kept current. It covers what carries over between sessions too. Prompt writing is one part of the larger job of building a reliable information environment for the assistant.
Rewriting in plainer words…
This answer doesn't lend itself to a diagram - it reads best . No credits were charged.
Context engineeringchooses, orders and trims the window
Instructionspicks
Chat historysummarizes
Files and docsfetches
Tool resultstrims
Prompt engineeringincludes
picks → Instructions
summarizes → Chat history
fetches → Files and docs
trims → Tool results
includes → Prompt engineering
Prompt engineeringthe words in one request
Promptwrites
Context window
Instructionsproject rules for this task
Chat historyearlier turns of this conversation
Files and docsfetched only when needed
Tool resultswhat tools send back
Promptone small piece
Current interfacecannot make up for
Prompt cannot make up for
Current interfacethe model cannot see it, so it guesses
Prompt engineering→Prompt(writes)
1step
From Context engineering to Instructions: picks.
1 / 7← → keys
The prompt is one small piece of what the model gets. Context engineering decides everything in the window, and no prompt can make up for what it leaves out.
Context engineering is designing everything the model sees during a task, while prompt engineering is about the wording of one request.
That context covers instructions, conversation history and summaries, source files, tests, docs, and whatever tool results come back.
It matters because a strong prompt can't fix missing context. If the model can't see the current interface or a project rule, it has to guess.
So the job is to put the right information in front of the model at the right time, and to cut noise that would distract it.
It also decides how context gets retrieved, ordered and kept current, and what carries over from one session to the next.
Prompt writing still matters, but it's one part of building a reliable information environment, and more context hurts when most of it is noise.
Weak answers to avoid
Calls it a new name for prompt engineeringSaying it just means longer prompts misses the difference in scope. Say it covers files, tool results, history and memory, beyond the wording of one request.
Assumes more context is always betterDumping the whole repo in adds noise that distracts the model. Say the job includes choosing and removing context, as well as adding it.
Lists context sources but never says why they matterA list alone misses the point. Explain that a model that can't see the current interface or a project rule has to guess, and no prompt fixes that.
Ignores freshness and carry-over between sessionsStale docs or old summaries mislead the model. Mention how context is retrieved, ordered and kept current, and what carries over between sessions.
Interview lens
Likely follow-ups, what you can say, and the weak answers to avoid.
