Write about what worked what didn't work. (language choice, open-source llm repo choice, challenges faced during this sprint and the previous one)