Numan builds AI health coach Nu with evaluation first
What happened
Numan built its AI health coach Nu on LangGraph and uses LangSmith to evaluate and trace conversations, limiting the service to lifestyle coaching and health education. The team evaluates single-turn replies by category, runs tests through the same agent graph as production, and blocks failing changes with safety regression tests. The same dataset let the team switch model providers within a day. An evaluation layer that simulates patient conversations dynamically is still a prototype, and criteria drafted by engineers need clinical sign-off before they can gate releases or back safety claims.
Written by AI from the coverage · updated 3 hours ago
Coverage
Follow the reports to see the story from different sides.
- AI HOT · Tips & opinionsNuman 如何用 LangGraph 和 LangSmith 构建评测优先的 AI 健康教练 Nu
Numan 用 LangGraph 构建 AI 健康教练 Nu，并通过 LangSmith 评测与追踪对话，将服务限定在生活方式辅导和健康教育范围内。团队按类别评测单轮回复，让测试经过与生产相同的智能体图，并以安全回归测试阻止不合格变更发布；同一套数据集支持其在一天内完成模型供应商迁移。动态模拟患者对话的评测层仍处于原型阶段，其工程师起草的标准须经临床专业人员签核后，才能用于发布门槛或安全声明。
Heat over time
Heat now 9·Comparable peak 9（Oct 20 01:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.