Comparison questionMedium~10 minAgents & Workflows / 8 of 18

Interview question

Compare context, state, and memory in an AI agent.

Separates the current model input from authoritative workflow state and selectively retained information across runs.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

Context is the bounded information sent to the model for the current step. State is the authoritative execution record, such as the goal, completed actions, approvals, results, and terminal status. Memory is selected information retained across turns or runs to improve future behavior. The application decides what enters each layer and what authority it has.

Full answer below

Practice focus

Agent data model, authority, persistence, context construction, and lifecycle.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Overview
These terms are often mixed together, but they have different authority and lifecycle.
Context is the bounded information sent to the model for the current step. State is the authoritative execution record, such as the goal, completed actions, approvals, results, and terminal status. Memory is selected information retained across turns or runs to improve future behavior. The application decides what enters each layer and what authority it has.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
A tool receipt belongs in state; a compact summary of earlier conversation may enter context; a user's stated formatting preference may become memory. Generated summaries and memories are derived and can be wrong, so they do not replace original events or verified facts. Context is rebuilt from relevant state and memory under token, privacy, and tenant rules rather than growing forever.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
Each layer has separate storage, ownership, versioning, retention, deletion, and concurrency rules. State transitions are durable and idempotent enough to resume after failure. Memory has provenance, confidence, scope, and expiry, and sensitive or consequential facts require verification before reuse. Context construction records which state and memory versions influenced an action so a failure can be reproduced without storing unnecessary hidden reasoning.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Calling the entire conversation history memory.
Treating a generated summary as authoritative state.
Persisting every observation indefinitely.

Red flagsCandidateRed flagsSignals that an interviewer may treat as risky.Open details

Details

Memory has no tenant or user scope.
A resumed run cannot recover completed actions.
No record explains which memory influenced a tool call.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

What belongs in authoritative state?
How would you compact context?
Which memories require verification?

Continue the section

Next in Agents & Workflows: Design agent memory

Step 9 of 18

How would you design long-term memory for an AI agent?

Practice route

Read core answer Review checklist Follow-up prompts Track progress

Finished this drill?

Practice freely. Sign in only when you want to save completion, confidence, notes, or bookmarks.

New to Aporeon? Create account

Drill details

Human editorial direction

Aporeon is shaped by Aleksandar Tomovski, a software developer with experience on both sides of technical interviews. Content is reviewed for accuracy, natural spoken delivery, useful depth, and honest trade-offs.

About the creator Editorial standards

Was this drill useful?

Help improve the interview library.

Found this answer useful?

Say thanks with a standalone, one-time $5 contribution. No account required.

No subscription or perks.Choose another amount

Opening drill

Comparison questionMedium~10 minAgents & Workflows / 8 of 18

Interview question

Compare context, state, and memory in an AI agent.

Separates the current model input from authoritative workflow state and selectively retained information across runs.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

Full answer below

Practice focus

Agent data model, authority, persistence, context construction, and lifecycle.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Overview
These terms are often mixed together, but they have different authority and lifecycle.
Context is the bounded information sent to the model for the current step. State is the authoritative execution record, such as the goal, completed actions, approvals, results, and terminal status. Memory is selected information retained across turns or runs to improve future behavior. The application decides what enters each layer and what authority it has.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
A tool receipt belongs in state; a compact summary of earlier conversation may enter context; a user's stated formatting preference may become memory. Generated summaries and memories are derived and can be wrong, so they do not replace original events or verified facts. Context is rebuilt from relevant state and memory under token, privacy, and tenant rules rather than growing forever.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
Each layer has separate storage, ownership, versioning, retention, deletion, and concurrency rules. State transitions are durable and idempotent enough to resume after failure. Memory has provenance, confidence, scope, and expiry, and sensitive or consequential facts require verification before reuse. Context construction records which state and memory versions influenced an action so a failure can be reproduced without storing unnecessary hidden reasoning.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Calling the entire conversation history memory.
Treating a generated summary as authoritative state.
Persisting every observation indefinitely.

Red flagsCandidateRed flagsSignals that an interviewer may treat as risky.Open details

Details

Memory has no tenant or user scope.
A resumed run cannot recover completed actions.
No record explains which memory influenced a tool call.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

What belongs in authoritative state?
How would you compact context?
Which memories require verification?

Continue the section

Next in Agents & Workflows: Design agent memory

Step 9 of 18

How would you design long-term memory for an AI agent?

Compare context, state, and memory in an AI agent.

Practice focus

Answer path

Overview

Review signals

Follow-up prompts

Continue the section

Compare context, state, and memory in an AI agent.

Practice focus

Answer path

Overview

Review signals

Follow-up prompts

Continue the section