LangChain NYC meetup to teach voice agent evaluation across execution, outcomes and experience

LangChain · x · 2026-09-10

LangChain is hosting a New York meetup on September 22, with software engineer Caroline di Vittorio presenting practical methods for evaluating voice agents.

The session covers three evaluation dimensions: execution (did the agent follow instructions and use tools correctly), outcomes (resolution, booking, transfer success), and experience (latency, interruptions, awkward pauses, conversational friction). It also covers when to use code evaluators, LLM judges, audio-aware evaluation and human review, plus tracing calls and building a continuous evaluation loop in LangSmith.

Original post →

More from coding & agent

coding & agent channel →