Prep RoomSign inStart free
Code RoomCache TTL intent mismatch
MediumPrep Room Coding #4506

Cache TTL intent mismatch

Vibe & agenticAlgorithms & data structuresMid–Senior~18 min

You asked an agent to add caching to an expensive API call in your Java service. Its first attempt cached with no expiry (stale data); you said 'add a TTL'; its second attempt added a TTL but caches per-request inside the request scope, so the cache never actually survives between calls. Each fix addresses your literal words but misses the intent. How do you re-steer, and when do you stop?

Implement
count_upstream_calls(keys: list[str], times: list[int], ttl_seconds: int) → int
Examples
in[["a","a"],[0,1],60]out1
in[["a","a","a"],[0,100,150],60]out2
in[["a","b","a"],[0,1,2],60]out2
What a strong answer looks like

Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.

0:00 of about 18 min

Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.

Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.