Skip to content
Claude Certified Architect — Professional

Anthropic · Topic 5 of 5

Evaluation, Quality, and Continuous Improvement

Measuring and improving quality over time: eval harnesses, LLM-as-judge, regression and canary/A-B testing, dataset curation, and the fine-tuning vs prompting vs distillation decision.

5 lessons150 questions
Start for freeFull certification

What this topic covers

  1. 01Building Eval Harnesses300m
  2. 02LLM-as-judge and Rubric Scoring300m
  3. 03Regression, Canary, and A/B Testing360m
  4. 04Dataset Curation and Feedback Loops360m
  5. 05Fine-tuning vs Prompting vs Distillation360m

Other topics in Claude Certified Architect — Professional

Advanced Agent Orchestration at Scale5 lessonsProduction Reliability and Observability5 lessonsSecurity, Safety, and Governance5 lessonsCost and Performance Engineering5 lessons