How Numan Built Evaluation-First AI Health Coach Nu with LangGraph and LangSmith
Original title:How Numan Built an Eval-First AI Health Coach with LangGraph and LangSmith
AI summary
UK digital health company Numan reviewed the evaluation system for its AI health coach Nu, concluding that the scarce asset is not prompts but evaluation criteria. Single-turn static datasets and scripted multi-turn dialogues both struggled to cover real conversations, so Nu switched to topic datasets scored by category, a simulated dialogue layer with cheaper models playing patients, and automatic blocking of releases on safety regressions.
Key Takeaways
- UK digital health company Numan reviewed the evaluation system for its AI health coach…
- Single-turn static datasets and scripted multi-turn dialogues both struggled to cover real…
- a simulated dialogue layer with cheaper models playing patients,and automatic blocking of releases…
Quick Facts
- Published
- Updated
- Source
- LangChain:Blog(RSS)
- Category
- Tutorials & Tips
- Topic
- Tutorials & Practice
- AI score
- 61
Numan 一手复盘 AI 健康教练的评测体系建设,给出的固定数据集、模拟对话和临床医生签核准则等做法可迁移到其他受监管场景。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: How Numan Built Evaluation-First AI Health Coach Nu with LangGraph and LangSmith. October 19, 2026. https://new.nat6.net/en/news/nk19eo4eud45d47tmwxm684iu/