Skip to content
AI Engineering Journal

Aide Hub

Practical notes on AI assistants, software architecture, and developer tooling.

Posts

1056

Tags

715

Featured

0

Latest

把 LLM Judge 放进 Agent 运行时

LLM judge 正从离线评测走进 agent 运行时,开始影响继续、重试、换模型或转人工。本文拆解 7 种架构模式,说明如何拆分判断、比较候选、检查过程,并用确定性逻辑控制风险。

Continue reading

Top Categories

View all tags

Trending Tags

Recent Posts

Browse archive