Skip to content

Tracks Drills Topics Technologies Tech stacks Skill areas

Tracks Drills Topics

Loading

Aporeon

Technical interview practice, from one developer to another.

About Editorial standards Support Privacy Terms Refund policy

Agents & Workflows - AI Engineering | Aporeon

Agents & Workflows

Tool use, planning, state, memory, permissions, approval, recovery, and agent evaluation.

18 drillsPractice one at a timeOrdered pathComparison questionConcept questionArchitecture scenarioDebugging scenario

Recommended start

Agent vs deterministic workflow

Distinguishes model-directed execution from explicit application control and explains when each is appropriate.

Medium~10 minStep 1 of 18

Practice this drill Choose another

Search

18 of 18 shown

1Agent vs deterministic workflowRecommended startDistinguishes model-directed execution from explicit application control and explains when each is appropriate.Comparison questionMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

Practice order

18 drillsOne at a time

1Agent vs deterministic workflowUp next / Medium / ~10 min
2When not to use an agentMedium / ~7 min
3Choose an agent workflow patternHard / ~15 min
4

2

When not to use an agentRejects agent complexity when rules, APIs, search, ordinary automation, or a bounded model call solve the task more safely.Concept questionMedium~7 minArtificial IntelligenceGenerative AILarge Language Models

Practice

3Choose an agent workflow patternChooses among routing, sequential, parallel, evaluator-optimizer, and bounded agent loops from the task's dependency structure.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

4Curate an agent toolsetBuilds a small, task-relevant, typed toolset with clear semantics rather than exposing an entire internal API surface.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

5Read vs action toolsClassifies tools by consequence and applies distinct permissions, validation, confirmation, idempotency, and audit controls.Comparison questionHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

6Bound an agent loopDefines success, failure, step, time, token, cost, retry, and repetition limits with useful escalation behavior.Architecture scenarioHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

7Plan and replan safelyUses plans as bounded, revisable execution aids while preserving evidence, policy, and application-owned state transitions.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

8Context vs state vs memorySeparates the current model input from authoritative workflow state and selectively retained information across runs.Comparison questionMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

9Design agent memoryStores only useful, scoped, attributable memories with extraction, verification, retrieval, expiry, correction, and deletion.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

10Human approval for agent actionsPauses durable execution with a clear action preview, verifies approver authority, handles expiry and changed state, and resumes idempotently.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

11Agent delegation patternsChooses delegation patterns based on who owns the user interaction, context, final answer, permissions, and failure handling.Comparison questionHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

12What MCP solvesDefines MCP as a client-host-server protocol for discovering and using tools, resources, and prompts without confusing interoperability with safety.Concept questionMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

13Secure remote MCPEstablishes server trust, audience-bound authorization, least privilege, capability filtering, approval, and safe handling of untrusted results.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

14API tools vs computer usePrefers typed APIs for reliable actions and reserves UI automation for constrained cases with stronger observation, approval, and recovery.Comparison questionHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

15Recover a durable agent runCheckpoints authoritative state, reconciles ambiguous tool outcomes, preserves approvals, and resumes without repeating completed work.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

16Evaluate an agentEvaluates verified outcomes, trajectory quality, safety, efficiency, recovery, and important slices in realistic environments.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

17Diagnose a looping agentUses traces to separate unclear goals, bad tool interfaces, missing state, contradictory results, poor termination, and retry amplification.Debugging scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

18Design a bounded agentCombines restrained autonomy, curated tools, durable state, approval, memory, recovery, MCP trust, evaluation, and operational controls.Architecture scenarioHard~20 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

Curate an agent toolset

Medium / ~10 min

5Read vs action toolsHard / ~12 min

6Bound an agent loopHard / ~12 min

7Plan and replan safelyHard / ~15 min

8Context vs state vs memoryMedium / ~10 min

+10 more drills in this section