AI engineering glossary · Agents tools MCP
What is evaluator-optimizer pattern?
Evaluator-optimizer pattern: A loop in which one model call produces a result and another evaluates it against criteria and gives feedback, repeating until the result is good enough. It works when quality can be judged more easily than produced.
Evaluator-optimizer pattern: an example
One call writes a product description; another checks it against a brand style guide and asks for revisions until it passes.
Related terms
More Agents tools MCP terms
- Fallback
- Generative agents
- Grounding (agents)
- Guardrails for agents
- Human-in-the-loop
- Human-on-the-loop
- Idempotency (tool calls)
- Indirect prompt injection
Where you learn it
Evaluator-optimizer pattern is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.