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.
66 results across 1 active filter
Page 1 of 3
Uses traces and saturation evidence to locate retry amplification, queue growth, exhausted workers, and failing fallback paths.
Diagnoses an AI regression by decomposing request stages, configuration changes, token growth, retries, routing, tools, retrieval, caching, and provider behavior.
Compares managed model APIs with operating open-weight models across control, expertise, privacy, scale, latency, cost, and portability.
Chooses delegation patterns based on who owns the user interaction, context, final answer, permissions, and failure handling.
Chooses between a direct provider relationship and a cloud-managed integration using concrete security, operations, capability, and commercial constraints.
Separates the current model input from authoritative workflow state and selectively retained information across runs.
Combines restrained autonomy, curated tools, durable state, approval, memory, recovery, MCP trust, evaluation, and operational controls.
Connects product success, layered evals, calibrated graders, release gates, guarded rollout, traces, monitoring, feedback, and incident learning.
Combines governed ingestion, authorized hybrid retrieval, reranking, grounded generation, citations, evaluation, observability, and fallback.
Combines threat modeling, injection containment, data minimization, tenant isolation, tools, audit, incident response, and governance.
Combines explicit trust boundaries, provider access, state, execution modes, resilience, budgets, versioning, and operations into a restrained design.
Abstracts stable application needs while exposing valuable capabilities deliberately instead of enforcing a false universal model API.
Uses plans as bounded, revisable execution aids while preserving evidence, policy, and application-owned state transitions.
Separates model and provider outcomes, applies bounded recovery, validates outputs, and exposes useful status and telemetry.
Treats prompts as observable behavior, removes secrets and authorization policy from them, and limits the consequence of disclosure.
Uses deadline budgets, transient-only bounded retries, jitter, idempotency, and propagated cancellation without retry storms.
Chooses a model from task-specific evaluation, operating constraints, safety, latency, cost, context, modality, and provider requirements.
Chooses among routing, sequential, parallel, evaluator-optimizer, and bounded agent loops from the task's dependency structure.
Selects relevant, trustworthy, and budgeted context instead of treating the context window as storage.
Bounds expensive inputs, outputs, retrieval, tools, retries, concurrency, and fan-out while preserving fair service.
Turns a product outcome into observable quality, safety, and operational criteria before selecting convenient scores.
Combines bounded instructions, source context, a narrow schema, abstention, semantic validation, review, observability, and rollout.
Designs a bounded and reversible first AI feature with measurable value, controlled data, validation, evaluation, observability, fallback, and human oversight.
Defines a narrow capability with a clear name, decision-oriented description, typed inputs, and safe result semantics.