Trace Every LLM Request Like Production Would

Replace ad-hoc logging with real structured request tracing.

A intermediate AI Engineering challenge worth 10 points. Solve it hands-on in a real AI Engineering environment in your browser - no local setup, no fake shells.

The Challenge

Scenario: The only record of what an LLM service is doing is a character count and a half-written log line - no request ID, no token count, no cost, no latency.

Integrate a structured JSONL tracer that records a request ID, model name, token counts, latency, status, and an estimated cost for every call, then query the trace file to prove the evidence is really there. Run check.