Pick a focused question that fits your time, stack, and interview goal.
How much time do you have?
Show one-drill sessions you can finish now.
160 results across 1 active filter
Page 6 of 7
Distinguishes exact, provider, semantic, and intermediate caches while preserving identity, freshness, versioning, and correctness.
Explains the operating, security, governance, and integration capabilities that surround managed model inference without treating a platform as the application.
Defines hallucination as unsupported or incorrect model output and connects it to evidence, task design, evaluation, and user-visible uncertainty.
Explains RAG as a runtime retrieval boundary that supplies relevant, current, and governed evidence without changing model weights.
Describes the context window as the bounded request state containing instructions, conversation, retrieved evidence, tool results, and output capacity.
Defines a large language model through next-token prediction and connects that mechanism to useful language behavior and its limits.
Defines AI evaluation as repeatable measurement of application behavior while preserving deterministic tests for code and contracts.
Defines embeddings as model-produced vectors used for similarity and contrasts representation with text generation.
Separates what a model can perform from whether the complete AI feature behaves dependably under real inputs, failures, and constraints.
Distinguishes model-directed execution from explicit application control and explains when each is appropriate.
Treats a production prompt as a versioned request contract containing instructions, context, output expectations, and application-owned controls.
Designs index records that support filtering, authorization, freshness, citations, diagnosis, and safe reprocessing.
Applies parsing, schema and domain validation, authorization, confirmation, idempotency, limits, and audit before execution.
Defines MCP as a client-host-server protocol for discovering and using tools, resources, and prompts without confusing interoperability with safety.
Places provider guardrails as one configurable defense while preserving application authorization, validation, testing, and incident evidence.
Treats model, prompt, schema, tools, routing, retrieval, policy, and evaluators as one observable behavior release.
Connects application, model, retrieval, tools, validation, retries, and outcomes while controlling sensitive telemetry.
Chooses the smallest execution model that satisfies the task while recognizing when explicit multi-step coordination adds real value.
Prefers typed APIs for reliable actions and reserves UI automation for constrained cases with stronger observation, approval, and recovery.
Chooses request-response, streaming, or durable background execution based on duration, user interaction, reliability, and side effects.
Uses pre-filtering for authorization and hard constraints while recognizing engine-specific recall and performance behavior.
Rejects agent complexity when rules, APIs, search, ordinary automation, or a bounded model call solve the task more safely.
Identifies tasks better served by deterministic code, search, rules, or conventional models because of correctness, latency, cost, or control.
Balances one-call simplicity against explicit multi-step workflows with intermediate validation and failure handling.