Solution Design and Architecture
Free Claude Certified Architect - Professional practice — 6 questions on Solution Design and Architecture, with explanations. No sign-up. Full 12-question mixed test →
A global enterprise processes 50,000 customer support tickets per day. Most tickets are simple account or billing questions, a minority require multi-step investigation across internal systems, and a small fraction are ambiguous and need a human. The company wants to minimize cost per ticket while keeping response quality high at this scale. Which architecture BEST meets these requirements?
At high volume, matching model capability to task complexity via a deterministic classify-then-route workflow minimizes cost while preserving quality where it matters, and reserves human escalation for genuinely ambiguous cases. This is the canonical cost/quality-tiering pattern for enterprise-scale support.
A financial services client needs a real-time trading dashboard chatbot that answers simple factual lookups with sub-second response times. Which model and API configuration is BEST suited to this requirement?
Haiku is the fastest model in the family for low-latency, low-complexity lookups, and streaming returns tokens as they're generated, minimizing perceived latency. Extended thinking is unnecessary for simple factual answers and would only add latency.
A RAG pipeline sends a static 20,000-token system prompt containing company policies on every request, and the service handles thousands of requests per hour. Which configuration BEST reduces cost and latency per request without losing information?
Prompt caching is specifically designed for this scenario: a large, unchanging prefix reused across many requests. Caching the policy block avoids reprocessing the full 20k tokens each time, cutting both cost and time-to-first-token on cache hits.
In Anthropic's documented agentic workflow patterns, what is the KEY distinction between the orchestrator-workers pattern and the evaluator-optimizer pattern?
This matches Anthropic's published agentic pattern definitions: orchestrator-workers dynamically breaks down and delegates subtasks based on the input, whereas evaluator-optimizer pairs a generator and a critic/evaluator in an iterative feedback loop to refine output quality.
During peak enterprise traffic, an application begins receiving intermittent 529 'Overloaded' errors from the Claude API. What is the BEST resilience pattern to implement in response?
529 errors are transient overload signals. The standard resilience pattern is exponential backoff with jitter to avoid retry storms, combined with a fallback path (different model tier, region, or provider) if retries are exhausted, preserving availability for end users.
A legal team wants an assistant that researches case law across several internal databases, decides which sources to query, synthesizes findings over an unknown and variable number of steps, and produces a final memo. The number of steps and sources needed differs significantly case by case. Which architecture BEST fits this requirement?
Open-ended tasks with an unpredictable number of steps and dynamic tool/source selection are the canonical use case for autonomous agents, as opposed to workflows, which are better suited to predictable, well-defined paths. The agent's own planning loop adapts to each case's variable complexity.
Ready for the real thing?
The full course has two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed answer explanations.
Start my full course on Udemy →