Skip to content

Tracks Drills Topics Technologies Tech stacks Skill areas

Tracks Drills Guides

Loading

Aporeon

Technical interview practice, from one developer to another.

Guides About Editorial standards Support Privacy Terms Refund policy

Prompting & Structured Output

Instruction hierarchy, context construction, schemas, validation, tool calls, and output repair.

20 drillsPractice one at a timeOrdered pathConcept questionArchitecture scenarioComparison questionDebugging scenario

Recommended start

Prompt as a contract

Treats a production prompt as a versioned request contract containing instructions, context, output expectations, and application-owned controls.

Easy~5 minStep 1 of 20

Practice this drill Choose another

Search

20 of 20 shown

1Prompt as a contractRecommended startTreats a production prompt as a versioned request contract containing instructions, context, output expectations, and application-owned controls.Concept questionEasy~5 minGenerative AILarge Language ModelsPractice 2Instruction hierarchyExplains instruction authority while keeping provider-specific message roles separate from the application’s durable trust boundary.Concept questionMedium~7 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 3Separate prompt inputsBuilds a readable request with stable instructions and clearly delimited, provenance-aware untrusted inputs.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 4Construct useful contextSelects relevant, trustworthy, and budgeted context instead of treating the context window as storage.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 5Zero-shot vs few-shotUses examples to clarify ambiguous task behavior while accounting for token cost, bias, maintenance, and evaluation evidence.Comparison questionMedium~7 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 6Decompose a model taskBalances one-call simplicity against explicit multi-step workflows with intermediate validation and failure handling.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 7JSON vs structured outputSeparates prompt-requested JSON, parseable JSON modes, and provider-enforced schema adherence.Comparison questionEasy~7 minGenerative AILarge Language ModelsPractice 8Design an output schemaDesigns a narrow, typed, provider-supported schema that communicates the real downstream contract.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 9Schema-valid but wrongSeparates structural validity from factual accuracy, evidence, authorization, and domain invariants.Concept questionMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 10Represent unknown valuesUses explicit status and field semantics so absence, null, not-applicable, and insufficient evidence are not conflated.Architecture scenarioMedium~7 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 11Evolve output schemasTreats the model-facing schema and downstream consumer contract as a versioned API with compatibility and rollout concerns.Architecture scenarioHard~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 12Output vs tool callingDistinguishes a typed final answer from a model request for the application to execute an external capability.Comparison questionMedium~7 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 13Design a tool contractDefines a narrow capability with a clear name, decision-oriented description, typed inputs, and safe result semantics.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 14Execute tool calls safelyApplies parsing, schema and domain validation, authorization, confirmation, idempotency, limits, and audit before execution.Architecture scenarioHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 15Return tool resultsReturns correlated, minimal, typed, and safely bounded tool results without confusing them with trusted instructions.Concept questionHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 16Handle incomplete outputSeparates provider refusals, safety blocks, token truncation, transport failure, and valid domain abstention.Architecture scenarioMedium~10 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 17Repair invalid outputUses constrained output where possible, deterministic parsing and validation, bounded targeted repair, and safe fallback.Debugging scenarioHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 18Stream structured outputSeparates user-visible progress from validated completion and prevents partial JSON or tool arguments from triggering side effects.Architecture scenarioHard~12 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 19Diagnose output qualityDiagnoses schema-valid quality regressions across inputs, context, prompts, examples, models, schemas, tools, and downstream interpretation.Debugging scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice 20Reliable AI extractionCombines bounded instructions, source context, a narrow schema, abstention, semantic validation, review, observability, and rollout.Architecture scenarioHard~15 minArtificial IntelligenceGenerative AILarge Language ModelsPractice

Practice order

20 drillsOne at a time

1Prompt as a contractUp next / Easy / ~5 min
2Instruction hierarchyMedium / ~7 min
3Separate prompt inputsMedium / ~10 min
4Construct useful contextMedium / ~10 min
5Zero-shot vs few-shotMedium / ~7 min
6Decompose a model taskMedium / ~10 min
7JSON vs structured outputEasy / ~7 min
8Design an output schemaMedium / ~10 min

+12 more drills in this section