Abstract: With the increasing complexity, scale, and adoption of GENAI workloads , it is crucial to consider their environmental impact. This chalk talk provides environmental sustainability insights that are specific to large language models and how machine learning operations MLOPS is critical to sustainable optimization. In particular, we discuss practical best practices for different customization scenarios including training models from scratch, fine-tuning with additional data using full or parameter-efficient techniques, retrieval augmented generation (RAG) and prompt engineering.