Seeking Guidance on Building Eval Pipelines in Phoenix
Hey everyone! 👋 Ouss here, founder of Palete (We deploy AI agent use cases on top of whatsapp for B2B). I’ve been using Phoenix pretty heavily for the last ~12 months, mostly for tracing, and it’s been super valuable. I’m still struggling a bit with building solid eval pipelines on top of it though. I’d love guidance on either:
clear resources / examples for eval workflows, or
chatting with someone who’s already doing this well
If anyone has pointers, repos, or is open to a quick discussion, I’d really appreciate it 🙏
