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.
45 results across 1 active filter
Page 1 of 2
Builds a hybrid retrieval pipeline that enforces tenant scope before ranking, assembles a bounded context, and verifies returned citations.
Separates application instruction attacks from attempts to bypass a model's safety behavior and explains their overlapping defenses.
Handles a poisoned retrieval corpus by freezing ingestion, tracing provenance, switching immutable index versions, rebuilding clean data, and proving recovery.
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.
Treats prompts as observable behavior, removes secrets and authorization policy from them, and limits the consequence of disclosure.
Returns correlated, minimal, typed, and safely bounded tool results without confusing them with trusted instructions.
Records attributable versions, decisions, evidence references, policy results, and side-effect receipts while minimizing content-rich logs.
Bounds expensive inputs, outputs, retrieval, tools, retries, concurrency, and fan-out while preserving fair service.
Uses provenance, content isolation, constrained authority, behavioral checks, and deterministic enforcement to contain hostile external instructions.
Designs a bounded and reversible first AI feature with measurable value, controlled data, validation, evaluation, observability, fallback, and human oversight.
Combines product policy, input and output controls, context-aware handling, human escalation, appeals, and measurable safety quality.
Pauses durable execution with a clear action preview, verifies approver authority, handles expiry and changed state, and resumes idempotently.
Stores only useful, scoped, attributable memories with extraction, verification, retrieval, expiry, correction, and deletion.
Coordinates private endpoints, DNS, egress policy, dependent services, observability, and recovery without confusing network isolation with authorization.
Coordinates lifecycle rules across application state, provider storage, files, embeddings, caches, traces, evaluations, and backups.
Makes AI involvement, limitations, evidence, consequential decisions, correction, escalation, and accountability visible to users.
Carries source permissions into ingestion and query-time filtering while failing closed across retrieval, caching, and citations.
Separates storage from processing location, verifies provider and feature-specific retention, and minimizes sensitive inputs.
Keeps credentials in trusted executors and replaces broad reusable secrets with scoped references and short-lived authority.
Applies purpose limitation, selective retrieval, redaction, pseudonymization, and field-level policy before model calls.
Constrains source-to-sink combinations with least privilege, egress policy, argument checks, approval, and information-flow awareness.
Carries trusted tenant scope through retrieval, caches, memory, tools, traces, evaluations, and asynchronous execution.
Builds system-specific adversarial testing around realistic assets, identities, channels, tools, and measurable impact.