Boost Claude Code performance with prompt learning - optimize your prompts automatically with evals

🚀 Boost the performance of Claude Code with prompt learning 🧠 Repo with all the code here: https://github.com/Arize-ai/prompt-learning Prompt Learning, a new concept from Arize, uses English language feedback and meta prompting to boost the effectiveness of your prompts. This video covers how this technique can be used to improve your claude.md or other coding agent rules file by running coding tests against known problems with known solutions, running evals in AX or Phoenix to get English feedback, then feeding that into a meta prompt to generate rules for the agent. The flow: - Run your coding agent against problems from SWE Bench - Log the problem, agent generated solution, ground truth solution, and unit test pass or fail to AX - Run an eval in AX to validate the generated solution and get English language feedback, saving the scores from your experiment - Send the feedback to a meta prompt and update the coding agent rules - Repeat multiple times to iterate on the prompt Chapters: 0:00 - Intro 3:18 - What is prompt learning 6:23 - Using SWEBench to test your coding agent 10:21 - Before and after scores 13:47 - Applying this to your own code 15:31 - Overfitting Learn more: Blog post on Prompt Learning - https://arize.com/blog/prompt-learning-using-english-feedback-to-optimize-llm-systems/ Repo with all the code here: https://github.com/Arize-ai/prompt-learning Arize documentation: https://arize.com/docs/ax Sign up for Arize for free: https://arize.com/ Phoenix, Open Source LLM observability and tracing: https://phoenix.arize.com/ #PromptLearning #Evals