Part 8 of 11

Evaluation and Observability

The deep spine: LLM-as-judge and its failure modes, component evals for RAG and agents, regression gates in CI, online evaluation, tracing, monitoring, and the data flywheel.

Chapters
9
Hours
2
Difficulty
Intermediate to Advanced
  1. 8.0intermediate

    The eval taxonomy

    Four axes that decide which kind of evaluation answers your question: offline vs online, component vs end-to-end, reference-based vs reference-free, automatic vs judge vs human.

    15 min
  2. 8.1intermediate

    LLM-as-judge

    Build a second-LLM judge you can trust: binary pass/fail over Likert, prompt patterns, and the human-agreement loop you must run before deploying it.

    15 min
  3. 8.2intermediate

    Judge failure modes

    Position bias, length bias, self-preference, and style bias in LLM-as-judge pipelines, with the numbers and the mitigations that actually work in production.

    15 min
  4. 8.3intermediate

    Evaluating RAG and agents

    Component evals, tool-choice accuracy, trajectory evaluation, and the weakest-link pattern that tells you which piece of a compound system to fix.

    10 min
  5. 8.4intermediate

    Regression suites and CI gates

    Wiring evals into CI as a real merge gate: two-tier golden sets, differential thresholds over absolute floors, and how to keep flaky probabilistic tests trustworthy.

    10 min
  6. 8.5intermediate

    Online evaluation and A/B testing

    Why your offline eval lied about a 3-point lift, the implicit signals that don't, and the sample sizes LLM A/B tests actually need.

    15 min
  7. 8.6intermediate

    Tracing

    Spans for LLM calls, retrieval, and agent loops; the gen_ai.* OpenTelemetry conventions; why vendor-proprietary tracing locks you in.

    9 min
  8. 8.7intermediate

    Monitoring and dashboards

    Why LLM systems need cost, latency, quality, and usage on one screen; how to alert on quality regressions and cost spikes; and which observability platform to pick in 2026.

    10 min
  9. 8.8intermediate

    Feedback & the data flywheel

    How production traffic compounds into better evals and a better product, instead of becoming a treadmill of sprint-by-sprint bug fixes.

    10 min