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.
1225 drills fit a 20-minute session. Closest fits first.
Page 12 of 52
Chooses Vertex AI shared or reserved capacity using traffic shape, reliability, geography, cost, and workload prioritization.
Tunes evidence quantity and no-answer behavior from measured relevance rather than treating similarity scores as universal confidence.
Combines low-friction ratings, structured reasons, behavioral outcomes, review queues, and privacy-aware eval case creation.
Bounds expensive inputs, outputs, retrieval, tools, retries, concurrency, and fan-out while preserving fair service.
Covers allocation rate, collection frequency, latency symptoms, and safe production steps for GC-heavy services.
Uses file upload processing to practice bounded memory, stream ownership, temporary files, and retained buffers.
Covers chunking, bounded concurrency, scoped dependencies, progress, retries, and avoiding full-dataset memory spikes.
Builds a repeatable path from governed source documents to parsed, chunked, embedded, traceable, and replaceable index records.
Places validation at the right form level and manages asynchronous checks without races or noisy requests.
Covers queue-backed work, idempotency, retries, visibility timeout, poison messages, progress tracking, and choosing Storage Queues or Service Bus.
Makes AI involvement, limitations, evidence, consequential decisions, correction, escalation, and accountability visible to users.
Moves from a reproducible user interaction to render evidence, root cause, proportional repair, and field validation.
Turns LOH theory into an API investigation around large payloads, buffers, serialization, and concurrency.
Places each risk at the cheapest reliable test boundary and keeps a small set of browser journeys for integration confidence.
Creates a task-oriented model boundary that normalizes application outcomes without erasing useful provider capabilities.
Connects answer claims to retrieved evidence while preventing decorative citations and preserving source-level traceability.
Explains safe migration strategy for Azure-hosted .NET APIs, including expand-contract changes, locking risk, rollback, and deployment order.
Combines safe bindings, AOT, dependency discipline, CSP, Trusted Types, and reporting without treating one control as a complete defense.
Combines server and client instrumentation, OpenTelemetry, post-response work, correlation, and privacy-aware signals.
Owns a third-party widget through a DOM ref, effect lifecycle, minimal updates, and complete teardown.
Gives a practical incident flow for memory growth using metrics, traffic correlation, heap evidence, and safe mitigation.
Coordinates source changes, versions, tombstones, reconciliation, and query-time freshness without treating the index as authoritative.
Separates durable application state from the model's bounded request context while preserving ownership, concurrency, and deletion.
Applies purpose limitation, selective retrieval, redaction, pseudonymization, and field-level policy before model calls.