Hi everyone. I just caught up on the AI Build and learn session from yesterday. Fab! I have a YouTube video on how to build your first eval pipeline. This doesn't cover any frameworks but rather foundational concepts on how to think about doing evals (Building a golden test set, code assertions, llm judges, and drift detection). Do check it out and let me know if you found it useful. I am currently experimenting with an end-to-end RAG pipeline and building out evals at every stage. I am interested in seeing how different chunking strategies, and other design decisions affect performance both for retrieval and generation and figuring out the best way to evaluate systematically.

@quaint-agent-46517 I like the youtube video. very well made.

Thanks Kevin! @busy-church-73395