Questions asked were around ML algorithms, why XGboost over other boosting algos. RAG architecture. Conversation AI library. Orcestration pipeline. LLM optimization - Caching, async calls, token limit problems. How do you optimize long prompts. Validation layers between RAG pipeline.